8th EEEIC Conference

Wroclaw University of Technology

Brandenburg Technical University
Technical University of Ostrava
Technical University of Koice
University Carlos III of Madrid

May 10-13 2009

International Conference on Environment and Electrical Engineering 2009

8 EEEIC International Conference

on Environment
and Electrical Engineering

Karpacz, Poland
10-13. May 2009
Editors: H. Schwarz [Link]
co-sponsored by IEEE Poland

ISBN 978-3-940471-10-9

http:\\[Link]

organiser / veranstalter
Wrocaw University of Technology
Department of Electrical Engineering
pl. Grundwaldzki 13
PL 50 377 Wrocaw
Technical University of Cottbus
Department of Power Distribution and High Voltage Technology
Walther-Pauer-Strae 5
D 03046 Cottbus

The German National Library CIP

One of this publication is by the German National Library.
ISBN

978 3 940471 10 - 9

copyright
Dieses Werk einschlielich aller seiner Teile ist urheberrechtlich geschtzt.
Alle Rechte, insbesondere die der Vervielfltigung, Verbreitung und bersetzung liegen beim Verlage. Kein
Teil des Werkes darf ohne Genehmigung des Verlages, auch nicht fr Unterrichtszwecke, fotografiert oder
auch durch andere Verfahren reproduziert und verbreitet werden. Ebenfalls ist eine Einspeicherung und
Verarbeitung in EDV-Anlagen ohne Genehmigung des Verlages strafbar.
publisher / verlag
2009, Technical University of Cottbus, Informations-, Kommunikations- und Medienzentrum (IKMZ)
printing / druck
Technical University of Cottbus, Reprografie
comb-bound book / bindung
Buchbinderein Reinhard Kuagk Cottbus

8th EEEIC Conference

Wroclaw University of Technology
Brandenburg Technical University
Technical University of Ostrava
Technical University of Koice
University Carlos III of Madrid

May 10-13 2009

International Conference on Environment and Electrical Engineering 2009

8 EEEIC International Conference

on Environment
and Electrical Engineering

Karpacz, Poland
10-13. May 2009
Editors: H. Schwarz [Link]
co-sponsored by IEEE Poland

ISBN 978-3-940471-10-9

Program Committee
Chairman Harald Schwarz Brandenburg Technical University (Germany)
* Sahin Albayrak Technische Universitt Berlin (Germany)
* Carlos lvarez Ortega Carlos III University of Madrid (Spain)
* Hortensia Amars Duarte Carlos III University of Madrid (Spain)
* Chiara Boccaletti University of Rome La Sapienza (Italy)
* Holger Borowiak Siemens AG (Germany)
* Antonio Bracale Universit degli Studi di Napoli Federico II (Italy)
* Guido Carpinelli Universit degli Studi di Napoli Federico II (Italy)
* Manuel Prez-Donsin Vigo University (Spain)
* Iftikhar A. Durraniius German Arab Technology Group Jeddah (Saudi Arabia)
* Radomir Gono Technical University of Ostrava (Czech Republic)
* Dariusz Grabowski Silesian University of Technology (Poland)
* Jan Izykowski Wroclaw University of Technology (Poland)
* Przemyslaw Janik Wroclaw University of Technology (Poland)
* Jin Jiang University of Western Ontario (Canada)
* Juha Karvanen National Public Health Institute (Finland)
* Antoni Klajn Wroclaw University of Technology (Poland)
* Wlodzimierz Koczara Warsaw University of Technology (Poland)
* Fabio Leccese Universit degli Studi Roma Tre (Italy)
* Dirk Lehmann Brandenburg Technical University (Germany)
* Zbigniew Leonowicz Wroclaw University of Technology (Poland)
* Veleslav Mach Technical University of Ostrava (Czech Republic)
* Dusmanta Kumar Mohanta Birla Institute of Technology (India)
* Eduardo I. Ortiz-Rivera University of Puerto Rico-Mayaguez
* Waldemar Rebizant Wroclaw University of Technology (Poland)
* M. Jaya Bharata Reddy Birla Institute of Technology (India)
* Stanislav Rusek Technical University of Ostrava (Czech Republic)
* Tomasz M. Rutkowski Lab. for Advanced Brain Signal Processing, RIKEN (Japan)
* Paco Ruz Universidad Politcnica de Cartagena (Spain)
* Tarlochan S. Sidhu University of Western Ontario (Canada)
* Tomasz Sikorski IASE & Wroclaw University of Technology (Poland)
* Krzysztof Siwek Warsaw University of Technology (Poland)
* Toshihisa Tanaka Tokyo University of Agriculture and Technology (Japan)
* Pierluigi Tenca Ansaldo Sistemi Industriali S.p.A. (Italy)
* Jan Wilbert European Aeronautic Defence and Space Company EADS N.V
* Ying Xin Innopower Superconductor Cable (China)
* Liu Zhaoxu Electric Power Research Institute - EPRI (China)
Honorary Committee
Honorary Chairman Tadeusz Lobos Wroclaw University of Technology (Poland)
* Janusz Fleszynski Wroclaw University of Technology (Poland)
* Matthias Mller-Mienack Vattenfall Europe Transmission GmbH (Germany)
* Stanislaw Osowski Executive Committee IEEE Poland Section
* Marian Sobierajski Wroclaw University of Technology (Poland)
* Zbigniew Sroka Wroclaw University of Technology (Poland)

P4

The Eighth EEEIC International Conference on Environment and Electrical

Engineering, Karpacz, 10-13 May 2009

EEEIC 2009

Welcome from the Conference Chair

In the time of increased awareness about the environment problems by the public opinion and
intensive international effort to reduce emissions of greenhouse gases, as well increase of the
generation of electrical energy to facilitate industrial growth, the conference offers broad
contribution towards achieving the goals of diversification and sustainable development.
The scope of the Conference is to promote a Forum, where people involved with electrical power
systems may exchange their experiences and present solutions found for actual and future
problems. The conference offers prominent academics and industrial practitioners from all over
the world the forum for discussion about the future of electrical energy and environmental issues
and presents a base for identifying directions for continuation of research.
This year, the organizers from Wroclaw University of Technology and Brandenburg Technical
University received strong support from our new partners: Technical University of Ostrava
(Czech Republic), Technical University of Kosice (Slovakia) and Carlos III University of Madrid
(Spain). We gratefully acknowledge the support from sponsors.
The renewed interest in electrical engineering and energy topics in general contributes to the
revival of the industry and encourages us to announce the next EEEIC 2010 Conference which
will be held in May 2009 in Prague (Czech Republic).

Harald Schwarz
BTU Cottbus (Germany)

P5

Sponsors of EEEIC 2009

ZPUE S.A.
ul. Jdrzejowska 79c
29-100 Woszczowa, Poland
[Link]

W. Investment Group
Institute of Power Systems Automation, Ltd.
ul. Wystawowa 1
51-618 Wrocaw, Poland, [Link]
International Visegrad Fund
[Link]

Urzd Marszakowski Wojewdztwa Dolnolskiego

[Link]

P6

EEEIC 2009
Regular Papers
Disturbances correlated to compensation of reactive power in a system with wind generation estimation of transient
components parameters
Przemyslaw Janik, Zbigniew Leonowicz, Jacek Rezmer, Piotr Ruczewski, Zbigniew Waclawek, Hortensia Amaris, and
Carlos Alvarez
1
Electrical Energy Measurements for Rome LV Customers by Distributed Web-Server Instruments
Fabio Leccese (invited paper)
8
Optimal Capacitor Placement on Interconnected Distribution Systems in Presence of Nonlinear Loads Using Genetic
Algorithms
Ahmad Galal Sayed
13
Computer Relaying Software Reliability Evaluation Using Event Tree
Aditya Anil, Dusmanta Kumar Mohanta, Diptendu Sinha Roy, and Ajit Kumar Panda
19
Relationship between Production Capacity of Wind Power Stations and Needs of Ancillary Services
Vaclav Cerny, Andrea Fialova, Lucie Houdova, Petr Janecek, and Eduard Janecek
23
Extracting Electricity from Groundwater Flow - A New Environment Friendly Source of Energy Case Study: Iran
Majid Labbaf Khaneiki and Ali Asghar Semsar Yazdi
27
Fault Liability and Maintenance Cost Modeling of a Steam Turbine
Lenka Houdova, Libor Jelinek, and Eduard Janecek
32
Voltage Stability Analysis of Grid Connected Wind Generators
Nguyen Tung Linh
36
Organic solar cells based on ZnO nanowires layer
Jiri Podzemsky and Wolfgang Schade
41
Wind Turbine Cylinders with Spiral Fins
Radomir Gono, Stanislav Rusek, and Miroslav Hrabcik
45
Reliability Standards for Large Interconnections
Petr Horacek and Petr Havel
49
Shields of Electromagnetic Wave Based on Amorphous and Nanocrystalline Soft Magnetic Materials
Mariusz Ozimek, Dominika Gaworska-Koniarek, and Wieslaw Wilczynski
54
Impedance-Source Inverter-Based High-Power DC/DC Converter for Fuel Cell Applications
Mikhail Egorov, Dmitri Vinnikov, Ryszard Strzelecki, and Marek Adamowicz
57
Voltage division over the interrupter units of a high voltage circuit breaker without grading capacitors
Henryk Stuermer
61
Adjustable Speed Generation System For Wind Turbine Power Quality Improvement
Wlodzimierz Koczara, Grzegorz Iwanski, and Zdzislaw Chlodnicki
65
Current progress in ambient energy harvesting using piezoelectric materials and electroactive polymers
Pawel Zylka
69
Thin Film Adhesion Measuring of Conductive Silicon Rubber Depending on Power and Coating Time in Magnetic
Sputtering Method
Artur Napierala and Jan Ziaja
73
Electrical behaviour of current conducting silicon rubber
Artur Napierala
77
On Hilbert Transform and its application to assessment of electrical signal waveform distortion
Tomasz Sikorski and Piotr Ruczewski
79
Voltage Dependence of Hologram Generation Dynamics in Nematic Liquid Crystal Hybrid Panels
Agata Anczykowska, Stanislaw Bartkiewicz, Lech Sznitko, and Jaroslaw Mysliwiec
83

P7

Relevance of a Safety Factor for Wind Power Trading in Comparison with the Utilization of a Storage
Steve Vller and J.F. Verstege
87
The Probabilistic Integration of Demand-Side Load and Generation in a Representatntive Irish Distribution Network
Keith Sunderland and Michael Conlon
91
Planning of Ancillary Services Securing Power System Operation
Ondrej Novak, Tomas Strnad, Petr Horacek, and Josef Fantik
95
Biofuels- the power from plants
Grzegorz Pawel Maliga and Amar Mohan Patil
99
Guideline for the Investigation of PEM Fuel Cell Systems in Automotive Applications
Maik Heuer, Mathias Kaebisch, Guenter Heideck, and Zbigniew Styczynski
103
Wind Turbine Driven Doubly-Fed Induction Generator With Grid Disconnection
Chitti Babu B, Mohanty K B, and Poongothai C
107
Possibilities of Utilization of Energy Briquettes
Gabriel Borowski
110
Optimum location and sizing of passive filters in distribution networks using genetic algorithm
Mahdi Ghiasi Varzeghan, Vahid Rashtchi, and Sied Hadi Hoseini
114
Balancing Voltage Source Direct Torque Control of Induction Motor with Three level Inverter
Aberrahmane Berkani and Hacene Rezine
118
Remote Irrigation controller under photovoltaic energy
Mohamed Lassad Benzaoui
124
Implementation of Inter-area Angle Stability Prediction in Wide Area Control
Ahmed A. Daoud
127
Some aspects of technical, economical and wind simplified analysis for 18 MW wind farm in Poland
Waldemar Dolega
133
The aspects of combustion and co-combustion biomass
Grzegorz Pawel Maliga and Amar Mohan Patil
137
Control and Stability Analysis of Doubly Fed Induction during Voltage Sags
Kairous Djilali Djilali, B. Belmadani, and M. Benghanem
141
Application and Comparison of LQR and Robust Sliding Mode Controllers to Improve Power System Stability
Sahar Saffar Shamshirgar, Mohammad Golkhah, Hojat Rahmati, and Mohammad Ali Nekoui
145
A Survey of Solar Energy Power Systems
Adekunle Babatope Ajayi
151
Reducing the Short Circuit Levels in Kuwait Transmission Network. (A Case Study)
Mahmoud Gilany and Wael Al-Hasawi
155
Impact of Inverter-fed Multi-Motor Drives on the Quality of Electric Power in the Mains
Maciej Stanislaw Pawlowski
160
Modeling analysis and solution of Power Quality Problems
Mahesh Singh, Vaibhav Tivari, and Rajkumar Jhapte
164
Power System Real Time Simulation Testing Using IEC 61850
Yogesh Devidas Sonawane, Pankaj M. Patil, Naresh D. Sonawane, and Sayali Y. Sonawane
170
Economical Benefits by Contribution of Large Wind Farms to Voltage Control
Ahmad-Rami Al-Awaad
175

P8

Optical properties of thin films of polyazomethine with triphenylamine unit in the main chain prepared by spin-coating
method
Marcin Palewicz, A. Iwan, J. Doskocz, W. Strek, and D. Sek
179
A new approach to a solar cell problem
Bronislaw Swistacz
183
Proposed Techniques for Identifying Faulty Sections in Closed-Loop Distribution Networks
Wael Al-Hasawi and Mahmoud Gilany
185
Charcoals Used as Electrode Material in Supercapacitors
B. Szubzda and R. Kulinski
190
Wavelets for detection of voltage dips and micro interruptions
Zbigniew Leonowicz
193
Two Modified Methods for Harmonic and Flicker Measurement based on RWPC considering spectral leakage and edge
effects
Marjan Saadati and Seyed Saeidollah Mortazavi
196
Student Papers
Combustion Characteristics of Palm Wastes in Fluidized Bed Combustor
Rosyida Permatasari, Kang Kin Hui, and Mohammad Nazri Mohd. Ja'afar
200
Dynamic Brake Regulator and Protection System in Locomotive
Farzad Takbiri and Siamak Farshad
204
Lightning Phenomenon - Introduction and Basic Information to Understand the Power of Nature
Luke Staszewski
208
Signal Processing Techniques used in Power Quality Monitoring
Umar Naseem Khan
212
Optimization of the costs of road lighting systems.
Andrzej Makw
216
Contemporary high voltage gapless surge arresters
Rafal Kurnatowski
218
Evaluation of Chaotic Ferroresonance in Power Transformers Including Nonlinear Core Losses
Ataollah Abbasi, Mehrdad Rostami, Hamid Radmanesh, and Hamid Reza Abbasi
220
Primary Energy Market and Electric Market interaction
Nuno Soares Domingues
226
Measurements of electromagnetic compatibility using GTEM chambers
Marcin Dbowski
230
Distributed Generation and Power Quality
Umar Naseem Khan
232
Using Ultracapacitors for Saving Energy in Regenerative Braking in Hybrid Vehicles
Amar Mohan Patil and Gregor Pawel Maliga
236
Solar energy in Spain
Dariusz Szymanski
240
Single Phase STATCOM-Its Control Algorthim
Linju Jose
243
Transients phenomena in capacitive voltage transformers- distance protection point of view.
Rafal Kurnatowski
247
Power quality in automatic restoration systems using stand alone generators
Piotr Danielski and Marcin Dbowski
249

P9

Ocean Power: tidal and wave energy

Dennis Grzelak, Nico Brose, and Lennardt Schnemann
253
Overvoltage Protection Devices
Cezary Pawel Szafron
257
High-voltage direct current technology
Tomasz Drobik
260
Transient response of doubly fed induction generator under voltage sag using an accurate model
Alireza Abbaszadeh, Saeed Lesan, and Vahid Mortezapour
263
Reverse electrodialysis for power generation
Sameer Deshmukh
268
Microbial Fuel Cells to produce electricity
Sameer Deshmukh and Yogesh Khambia
272
Cable Shielding to Minimize Electromagnetic Interference
Anke Frbel
280
Fault Location System for Transmission Lines in One-terminal By using Impedance-Traveling Wave Assembled Algorithm
Sachin Ramrao Yerekar
283
Market behaviour to promote renewable energy
Nuno Soares Domingues
287
The future goes offshore
Robert Hoyer, Henry Rmer, and Jonathan Saudhof
293
Strong Electromagnetic Impulse Generation
Daniel Pyda
296
Detection of Fault Position with Respect to the Compensating Bank in Series Compensated Line by Measurement of Phase
Shift for Distance Relay Input Currents
Piotr Mazniewski and Jan Izykowski
300
Lightning Protection of Aircraft
Lukasz Mackowiak
304
Elimination of Chaotic Ferroresonance in Power Transformers Including Nonlinear Core Losses Applying of Neutral
Resistance
Ataollah Abbasi, Mehrdad Rostami, Hamid Radmanesh, and Hamid Reza Abbasi
308
Geothermal energy as a renewable energy source
Stefanie Kitscha, Jrgen Schmeckebier, and Marcel Steckel
314
Slot harmonics in stator current as syptoms of reparation of induction motors
Jan Stanislaw Pytlarz and Pawe Ewert
318
Model of a Wind Turbine with PMSG and Wind Park Configuration
Michal Piotr Jankow
321
Load Modeling For Power System Analysis
Jakub Kepka
325
Loads and Synchronous Generator Modeling
Robert Pliszczak
329
Renewable energy from the oceans Wave power, tidal power and ocean streams
Lars Bork, Bastian Garnitz, Thomas Krisch, and Carsten Schwartze
333
Carbon Capture and Storage Part of the solution to the climate change problem?
Robin Kuscheck, Sren Schlter, and Martin Noack
337
FFT algorithm optimization using Stream SIMD Extension instruction set
Pawel Dawidowski

P10

339
Abuse of transmission congestion existing naturally or created by bidding strategies in order to exert market power
Abouzar Estebsari, Seyyed Mohammad Sadeghzadeh, and Mehrdad Rostami
341
"Banning of Incandescent Bulbs and Its Influence to Electromagnetic Compatibility
Yan Lu
346
Reactive power in one-phase circuits with periodical voltages reactive power compensation
Sebastian Sabosz
349
Grid integration of renewable energy sources
P. Kammer and Andreas Kober
353
Solar potential of the Sahara Desert with an introduction to solar updraft power plants
Stefanie Fiedermann, Jadranka Halilovic, and Torsten Bogacz
357
Power Quality in distribution power networks with photovoltaic energy sources
Ricardo Albarracin Sanchez and Hortensia Amaris
360

P11

Disturbances correlated to compensation of

reactive power in system with wind generation
estimation of transient components parameters
P. Janik, Z. Leonowicz, J. Rezmer, P. Ruczewski, Z. Waclawek, H. Amaris*, C. Ortega*
Wroclaw University of Technology, Wybrzee Wyspiaskiego 27, 50-370 Wrocaw (Poland)
Electrical Engineering Dep., Universidad Carlos III de Madrid, Av. de la Universidad 30, 28911 Legans Madrid (Spain)
Abstractprecise computation of current components is
a key prerequisite for reliable assessment of power quality.
Especially in networks with wind generation we may
observe increased number of possible disturbing
phenomena. This paper presents an approach to accurate
computation of currents components with two similar
parametric methods based on singular value decomposition
(SVD) and Prony model. Those methods seem to be
applicable for the detection of non integer multiples of the
main frequency in decaying signals. Results of both methods
have been compared and evaluated. with respect to
traditional Fourier approach.
Index Termspower quality, transients, SVD, Prony
method.

I. INTRODUCTION
The widespread implementation of wind energy
conversion systems is a reality. Wind is seen as a clean
and renewable energy source, so the development of wind
generation technologies is welcomed and supported by
ecologists and governments. In the next years we will
have even more generator units connected to the grid [1].
Wind turbines, despite of their advanced control
systems and power electronic converters, influence in
many different aspects the electrical system they are
connected to [2, 3]. So far, the electrical distribution
networks were designed and operated under the
assumption of centralized generation and an energy flow
from the substation to the consumer. It is no more the
case [3, 4]. The connection of wind generators could lead
to many disturbances, such as: voltage fluctuations,
flickers, harmonics, instability, blind power regulation
problems, and transients [10]. Power quality issues
connected with wind generation are not only important
because of technical aspects, they are also crucial on the
free energy market.
There are at least three main wind generators
structures, which can be easily pointed out [4]. The
simplest and previously popular is the squirrel-cage
induction generator connected directly to the grid.
Usually, that type of turbines has a fixed pitch of turbine
blades. Second is the doubly-fed induction generator. The
stator winding of this generator is coupled with the

system grid, and the rotor winding is connected to a

voltage-source converter. The converter adjusts the
frequency of the rotor feeding current in order to enable
variable speed operation. The wind generator operates in
wide spectrum of wind speeds and has lower impact on
the grid, but the investment costs are higher. The third
structure of wind generation unit has a synchronous
machine. The rotating shaft and generator are coupled
directly without gear box. That generators type requires a
back-to-back converter for the grid connection, but it can
be operated in wide wind change range. Additionally,
voltage, active and reactive power can be controlled, as in
double feed induction generator.
Many of the wind energy converters installed today
still have a squirrel-cage induction machine connected
directly to the grid [6, 7]. This type of the generator
cannot perform voltage control and it absorbs reactive
power from the grid. Phase compensating capacitors are
usually directly connected. That type of wind turbine is
cheap and robust and therefore popular, but from the
system analysis point of view it has some drawbacks [4].
An important disadvantage is that during the switching
of the phase compensating capacitors, transients occur
[7], which can be disturbing for sensitive equipment,
protection relays and insulation. Also the impact on
power quality indices can not be neglected [5, 10].
Transient overvoltages can theoretically reach peak
values up to 2.0 pu. High current transients can reach
values up to ten times the nominal capacitor current with
a duration of several milliseconds [8].
The purpose of this paper is the assessment of
transients in electrical system for wind turbines equipped
with an asynchronous generator. A wind energy converter
connected to a distribution system was modeled in
Matlab SimPowerSystemsTolbox [9].
The Prony model and SVD method were considered as
appropriate tools for the parameters estimation of
transients. The analysis was carried out for different
operation conditions of the wind energy converter.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
1

X1(M) = SX 0(M)

II. COMPUTATIONAL METHOD FOR SPECTRUM

ESTIMATION BASED ON SINGULAR VALUE
DECOMPOSITION (SVD)

x(t ) = A1e + A2 e + ... + A e

s1t

s2t

s t

(1)

where: A - signal amplitude, s = + j - complex

frequency.
For practical reasons we consider only digital signals,
sampled with the interval T
(2)
t = [nT ]
The signal samples are given by

xn = A1 (es1T )n + A2 (es2T )n + ... + A (es T )n =

A1 z1n + A2 z2n + ... + A z n

(3)

The practical realization of the proposed method

imposes designing of a digital filter with finite impulse
response (FIR). This filter (Fig. 1) should block all signal
components. The filter transmittance in general form is
given by

H ( z ) = 1 + h1 z 1 + h2 z 2 + ... + h z
xn

(4)

yn

H(z)

Fig. 1. Basic digital filter scheme as a practical realization of proposed

analysis method

The filter response is given by

Y(z) = H(z) X(z)

(5)
The blocking property of the considered filter results

in

y = x + h1 x 1 + h2 x 2 + ... + h x0 = 0
yn = 0 for n

(6)

Equation (6) can also be written in a practice relevant

matrix form

x1 0
x 0
2
M = L

x N 1 0
x N h N

h N 1

h N2

0 x0
0 x1

L M

L 1 x N 2
L h1 x N 1
(7)

or

x1 = Sx0

X1(M) matrix is given in (10).

The construction of the

The proposed method, based on Singular Value

Decomposition (SVD) [14, 15, 16], is a technique for
modeling sampled data as a linear combination of
exponentials.
Accordingly, the signal x(t ) can be described
mathematically as a sum of exponential components:

The

(9)

matrix is constructed in a similar manner.

X(M)
0

X1( M )

x1
x
2
x3
=
L
x N 1

xN

xM

x2

x3

x M 1

x3

x4

xM

x M +1

x4

x5

x M +1

x M+2

xN

x N +1

x N + M 3

x N+ M2

x N +1

x N+2

L x N +M 2

x N + M 1

(10)
Similarly, equation (3) can also be given in a modified
matrix form

x1 = W0 ZA

(11)

where

1
z
1
z12
W0 =
L
z1N 2
N 1
z1

1
z1

z 22
L
z 2N 2
z 2N 1

z 32
L
z3N 2
z3N 1

z1 0 0
0 z
0
2

0 0 z3
Z=
L L L
0 0 0

0 0 0
A = [ A1

L
L

1
z1

1
z1

L z 2N 1
L L
L z NN 12
L z NN 11

L
L

0
0
0

L
0

z N

0
0

L 0
L L
L z N 1
L 0

A 2 A N-1

1
z1
z 2N
(12)
L
z NN 2

z NN 1

AN ]

(13)

(14)

Considering (8) and (11) the following equation can

be formulated

W0 ZA = SW0 A

(15)

finally

S = W0 ZW01

(16)

The latest equations shows, that z1 , z2 , z3 ,..., z are the

eigenvalues of the matrix S .
Regarding (9) the matrix S may be calculated
(M)T 1
S = X1(M) X(M)T
( X(M)
)
0
0 X0

Knowing the

z1 , z2 , z3 ,..., z

(17)

values the complex

(8)

frequencies (3) s1 , s2 , s3 ,..., s may be easily calculated.

Generally, the realization of proposed method implies

using M-column matrices instead of one column vectors
of input signal samples xn.

Similarly the values of amplitudes A1 , A2 , A3 ,..., A may

be derived from (11).

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
2

III. PRONY MODEL APPLIED FOR COMPUTATION OF

SPECTRAL COMPONENTS IN SIGNALS
The Prony method is a technique for modeling
sampled data as a linear combination of exponential
functions [8]. Although it is not a spectral estimation
technique, the Prony method has a close relationship to
the least squares linear prediction algorithms used for AR
and ARMA parameter estimation. Prony method seeks to
fit a deterministic exponential model to the data in
contrast to AR and ARMA methods that seek to fit a
random model to the second-order data statistics.
Assuming complex data samples, the investigated
function can be approximated by p exponential functions:
p

y[n] = Ak e

( k + jk )( n 1)Tp + j k

(18)

k =1

where
n = 1, 2,..., , T p - sampling period, Ak - amplitude,

k - damping factor, k - angular velocity, k - initial

phase.
The discrete-time function may be concisely expressed
in the form
p

y[n] = hk zkn 1

(19)
(

+ j )T

zk = e k k p
hk = Ak e j k ,
The estimation problem is based on the minimization
of the squared error over the N data values

= [ n]

(20)

n =1

where
p

[ n] = x [ n] y [ n ] = x [ n ] hk zkn 1 (21)
k =1

This turns out to be a difficult nonlinear problem. It

can be solved using the Prony method, that utilizes linear
equation solutions.
If as many data samples are used as there are
exponential parameters, then an exact exponential fit to
the data can be made.
Consider the p-exponent discrete-time function:

x[n] =

k =1

hk zkn 1

(24)

The polynomial may be represented as the sum:

F (z) =

a [ m]z p m =

m =0
p

= a [ 0] z + a [1] z

p 1

(25)

+ K + a [ p 1] z + a [ p ]

Shifting the index on (5) from n to n-m and

multiplying by the parameter a[m] yield:
p

a [ m ] x[n m] = a [ m ] hk zkn m 1

(26)

k =1

Equation (9) can be modified into:

p

a [ m] x[n m] =

m=0

(27)
p

p m 1
=
a [ m ] zk

k =1
m = 0

The right-hand summation in (10) may be recognized

as a polynomial defined by (8), evaluated at each of its
roots yielding the zero result:
p

hk zkn p

a [ m] x[n m] = 0

where

F ( z ) = ( z zk ) = ( z z1 )( z z2 )K z z p

k =1

Prony proposed to define the polynomial that has the

exponents as its roots:

(22)

(28)

m=0

The equation can be solved for the polynomial

coefficients. In the second step the roots of the
polynomial defined by (8) can be calculated. The
damping factors and sinusoidal frequencies may be
determined from the roots zk .
For practical situations, the number of data points
usually exceeds the minimum number needed to fit a
model of exponentials, i.e. > 2 p . In the over
determined data case, the linear equation (11) should be
modified to:
p

a [ m] x[n m] = e [ n]

(29)

m=0

The estimation problem is based on the minimization

of the total squared error:

E=

e [n]

(30)

n = p +1

k =1

The p equations of (5) may be expressed in matrix

from as:
z10
z20 K z 0p h x [1]

z11
z12 K z1p h2 x [ 2]
=
(23)

M
M M M
M

p 1
h
z2p 1 K z pp 1 p x [ p ]
z1
The matrix equation represents a set of linear
equations that can be solved for the unknown vector of
amplitudes.

IV. SIMULATION OF SYSTEM WITH INDUCTION

GENERATOR AND CAPACITORS
The wind generator with compensating capacitors is
shown in Fig.2. The simulation was done in Matlab using
the SimPowerSystem Toolbox [9].

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
3

Bus 1

Bus 2

Bus 3

Bus Sys

IM
Generator

Trafo

Line

source with short circuit capacity of 100 MVA and X/R

ratio of 7. The induction generator reactive power
demand varies with the produced real power [10]. During
the research different compensation levels were
simulated. Simulation results correlate with measured
values [7].
V. RESULTS OF SIGNAL ANALYSIS

Fig. 2. Induction generator with compensating capacitors.

A wind turbine generates power and accordingly a

mechanical torque on the rotating shaft, while the
electrical machine produce an opposing electromagnetic
torque [4]. In steady state operation, the mechanical
torque is converted to real electrical power and delivered
to the grid. The power generated by the wind turbine is
[4, 6]

P=

1
AC pV 3
2

(31)

The capacitors indicated in Fig. 2 were build as a bank

of separately closed units, what is close to practice [7]. In
this practical case two-stage phase-compensating
capacitors were connected at the wind turbines. One
small (C1=0.5 mF), for low rotor speed and both (C1+
C2=1.7 mF), when the wind turbines operated at higher
rotor speed.
Transients occurring during switching of the capacitor
bank are shown in Fig. 4. Clearly visible are two stages
of switching.

T=

Current [A]

and the torque can be found as

(32)

where - density of air, A - swept area of the blade,

C p - performance coefficient, V

- wind speed, T

mechanical torque, P - output power of the turbine, s

rotor speed of the turbine. At the constant wind speed, the
C p coefficient depends on the rotor speed s and pitch
angle and is often presented in a table form [10]. The
turbine characteristic used in simulation is shown in
Fig. 3.

Fig. 4. Currents during two steps of capacitor closing one phase.

A. Disturbances due to first capacitor switching

Power [pu]

The analysis of transients after the first capacitor are

analyzed in this chapter. More detailed waveform of the
current is shown in Fig. 5.

Current [A]

2000

1000

Fig. 3. Induction generator output power vs. angular velocity

The pitch control dynamic can be neglected in power

system transient analysis [6]. The simulated generator is a
150 kW, 400 V, 1487 rpm, induction machine. It is
connected to the grid through a Dyg 25/0.4 kV
distribution transformer which nominal power was varied
between 0.5 and 2 MW during the research process and
other parameters were set with accordance to [11]. A
typical 5 km overhead line [11] connected the generator
to a system. The system was represented by equivalent

-1000

0.02

0.04
Time [s]

0.06

0.08

Fig. 5. Transient current during the first step of capacitor closing.

Fourier analysis indicated two spectral components

(Fig. 6). However, the computation of amplitude of a
decaying component is not accurate using Fourier.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
4

160
50.00

SVD

Current [A]

49.99
120
80

587

40
0

0
0

500 1000 1500 2000 2500 3000

Frequency [Hz]

592

500 1000 1500 2000 2500 3000

Frequency [Hz]

Fig. 9. Two spectral components computed with SVD method

Fig. 6. Fourier transform of the current during first capacitor closing

Therefore Prony and SVD approach were used to

compute detailed signal parameters.

Current [A]

150

Similar results have been achieved with SVD.

Frequencies indicated by Prony are shown in Fig. 9.
Detailed information of the first and second signal
component obtained with Prony and SVD model are
given in Table I. The information includes the values of
amplitude A, decaying time constant , frequency f,
and initial phase .
TABLE I
CURRENT COMPONENTS AFTER FIRST CAPACITOR SWITCHING

Signal com.
[,o.]
1.
2.

-150

0.02

0.04
Time [s]

0.06

1.
2.

0.08

[s]
Prony
162.9
-2225
0.004
SVD
164,7
-2128,9
0.004
I
[A]

f
[Hz]

[rd]

49.86
592.2

0.87
1.5

49.43
588,2

0.90
1.52

Fig. 7. First component in the current, reconstructed with Prony method

Fig. 7 and 8 depict the main 50 Hz current component

and the switching transient reconstructed with parameters
given by Prony model.

B. Disturbances due to second capacitor switching

Similar procedure was applied for the transient current
after closing of the second capacitor (Fig. 10).
800

2000

Current [A]

Current [A]

400
1000

-400

-1000

0.02

0.04
Time [s]

0.06

0.08

Fig. 8. Second component in the current, reconstructed with Prony

method.

0.02

0.04
Time [s]

0.06

Fig. 10. Current after second capacitor closing

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
5

0.08

60
10

40

Current [A]

Current [A]

50

30
20

-10

10
0

500 1000 1500 2000 2500 3000

Frequency [Hz]

0.02

0.04
Time [s]

0.06

0.08

Fourier transform of the current during second capacitor

Fig. 14. Second component in the current, reconstructed with Prony

method.

Fourier transform (Fig. 11) of the current in Fig 10

indicates three components.

Similar results were obtained with SVD method (Fig. 15).

Fig. 11.
closing

SVD

Current [A]

40

347 352

-40

0
Fig. 12.
method

0.02

0.04
Time [s]

0.06

0.08

First component in the current, reconstructed with Prony

Fig. 13, 13 and 14 depicts the three components given

by the Prony model.

Current [A]

700

350

500 1000 1500 2000 2500 3000

Frequency [Hz]

Fig. 15. Three spectral components computed with SVD method

Detailed information of the three components obtaiend

with SVD and Prony method is given in Table II. The
information includes the values of amplitude A,
decaying time constant , frequency f, and initial
phase .

Signal com.
[,o.]

I
[A]

1.
2.
3.

51.21
13.84
831.0

1.
2.
3.

48.85
12.08
736,8

-350

504 510

49.96 50.00

0.02

0.04
Time [s]

0.06

0.08

[s]
Prony
-0.44
0.004
SVD
-0.47
0.004

f
[Hz]

[rd]

50.97
349.2
519.2

0.75
0.30
1.91

50.45
352,11
512.27

0.76
0.32
2.01

Fig. 13. Second component in the current, reconstructed with Prony

method.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
6

VI. CONCLUSIONS
The research results show, that the Prony model and
SVD method are useful for transient estimation in
systems with wind generators and compensating
capacitors. Those methods enabled accurate estimation of
amplitude, time constant, phase and frequency of
transients components of simulated signals. The current
waveform and its parameters depend on the capacitor to
be switched. Those dependences could be observed
during simulation.
Application of both Prony Model and SVD method
required a signal model. Fourier transform, as non
parametric method, did not require a signal model or even
the number of components, but could not compute signal
parameters besides frequency. Two signal components
were predefined, so the rest was considered noise.
The order of the signal model could be easily extended
to detect additional components. Both methods delivered
similar results, however, application of Prony model
seems more suitable for estimation of signal parameters.

REFERENCES
[1] Hammons T.: Status of Integrating renewable electricity
production in Europe into the grids. Proceedings of
Universities Power Engineering Conference, Cork
(Ireland), 2005, p.73
[2] Hanzelka Z., Mroz M., Pawelek R., Pitek K.: Quality
Parameters of 15kV supply voltage after connection of
wind farms-case study. Proceedings of Conference on
Harmonics and Quality of Power, Cascais (Portugal),
September 2006, on CD, paper ID 124
[3] Quinonez-Varela G., Cruden A., Grant A. D.,
Castaneda A.: Electrical Integration Assessment of Wind
Turbines into Industrial Power System: The Case of a
Mining Unit. Proceedings IEEE Power Tech Conference,
Porto (Portugal), September 2001, paper ID DRS3-292
[4] Chompoo-inwai Ch., Wei-Jen L., Fuangfoo P., Williams
M., Liao J. R.: System Impact Study for the Interconection
of Wind Generation and Utility System. IEEE Trnas. on
Industry Applications, vol. 41, January 2005, pp 163-168

[5] IEC 61400-21 Wind turbine generator systems. Part 21:

Measurement and assessment of power quality
characteristics of grid connected wind turbines.
[6] Tabesh A., Iravani R.: Transient Behavior of Fixed-Speed
Grid-Connected Wind Farm. Transactions of International
Conference on Power Systems Transients, Montreal
(Canada), June 2005, Paper No. IPST05-068
[7] Thiringer T.: Power Quality Measurments Performed on a
Low-Voltage Grid Equipped with Two Wind Turbines.
IEEE Trans. on Energy Conversion, vol. 11, September
1996, pp. 601-606
[8] Lobos T., Rezmer J., Kolin H.-J.: Analysis of Power
System Transients using Wavelets and Prony Method.
Proceedings IEEE Power Tech Conference, Porto
(Portugal), September 2001, paper ID EMT-103
[9] The Matworks: SimPowerSystems Users Guide. The Math
Works Inc., 2006
[10] Lubosny Z.: Wind Turbine Operation in Electric Power
Systems. Springer Verlag, Berlin Heidelberg (Germany),
2003
[11] SIEMENS: Electrical Engineering Handbook, 1987
[12] Bhatti A. M.: Practical OptimizationMethods. Springer
Verlag, New York (USA), 2000, pp. 288-302
[13] Larsson A. Thiringer T.: Measurment on and Modelling of
Capacitor-Connecting Transients on a Low-voltage Grid
Equipped with Two Wind Turbines. Proceedings IPST
Conference, Lisbon (Portugal), 1995
[14] T Lobos, T. Kozina, H.-J. Koglin: Power system
harmonics estimation using linear least squares method
and SVD,in IEE Proc., Generation Transmission and
Distribution, vol. 148, No. 6, 2001, p. 567-572,
[15] T. Will: Introduction to the Singular Value
Decomposition
Available:
[Link]
[16] L. Hexi, W. Guorong, Shi Y.; Z. Weimin, The Automatic
Rcognition of Welding Targets Based on ormalized SVD
of Image Matrix in Proc. International Conference on
Mechatronics and Automation, ICMA 2007 ,August 2007,
pp. 3100 - 3104

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
7

Electrical Energy Measurements for Rome LV

Customers by Distributed Web-Server Instruments
Fabio Leccese
Electronic Engineering Department, University of Roma Roma Tre
Via della Vasca Navele n.84, 00146, Roma, Italy, tel. +390657337085, fax +390657337101
leccese@[Link]
Abstract-Power Quality (PQ) measurements for low voltage
(LV) customers, domestic and business, have been carried out in
order to detect the behaviour of the electric net over time and the
quality of the electric energy being bought. In order to measure
PQ parameters, we realized an instrument based on web server
personal computers, which are common in office or in domestic
environment. This allows us to conjugate the high PC calculus
capability with the possibility to send data via internet to a central
server; moreover, the use of the existing hardware infrastructure
makes the instruments extremely cheap.

In this article we propose the results of a first measurements

campaign in the Rome area realized by means of instruments
distributed over the territory, inserted in LV plant, with the aim
to analyze the PQ parameters of power supply for LV
customers. After a discussion regarding the design choices for
the instrument, we will show the first monitoring results.
II.

THE LOCAL INSTRUMENT

The instruments are conceived to be cheap with the idea to

be easily placed in the final customers site in order to realize a
With the constant rise of raw materials prices and the net more and more widespread. The peculiarity of this
consequent increase of the power supply purchase costs, the instrument is the exploitation of an already existent web
necessity to evaluate the quality of this good is becoming more architecture for other aims. In fact, inside offices, shops or
and more important[1,2,3]. In fact, the low prices of the power houses, LV customers do own servers, which are always active
supply have, in the recent past, discouraged PQ analyses and constantly connected to the internet. We thought of using
leaving them only to professionals and experts that normally the high calculus capability of these PCs, often not adequately
work in the research environment. In the electrical market, only exploited, to locally elaborate the electrical energy samples
big companies have specialists, the energy managers, able to coming from an analogical interface connected to the PC audio
buy the energy needed for the proper functioning of business [4] card that acts as an A/D card. After the data acquisition and
with the idea to try to make the costs, related to consumptions, elaboration, the results are sent to a central server placed at the
matched with the quality of the energy purchased. Anyway the University of Roma Tre. Using existing PCs, it is obvious that
absence of continuative measurements carried out on the the costs are reduced, as compared to the analogical interface
electrical net makes impossible to evaluate the quality of and to the set up procedure.
electrical energy, forcing the companies to adopt alternative
Currently, the local instrument is composed of three
solutions to compensate the possible lack of quality. It is the principal parts:
case of big companies like Telecom Italia that, in order to a PC, which the customers already own and often used for
assure the continuity of the service, adopts complex and
other aims; for example as web server or net data storage
redundant electric plants, but also of LV customers that, for
system, always switched on and always connected to the
example, use UPS connected to their PCs. This is reflected into
internet with an ADSL line;
additional costs which are almost exclusively covered by the an analogical interface with sensors for current e and
customers[5].
voltage connected to the stereo line in of the PC audio
LV customers are particularly affected by this problem: both
card;
because their little commercial dimensions lower their the software capable of analyzing PQ parameters.
capability to bargain over the price of power supply and, above
A. The PC and the Audio Card Calibration
all because they cannot fully realize their needs and
It is not important the model of the PC, the only significant
expectations towards this good. This incapacity is strictly due
element
is that it must have an audio card. Therefore, it is
to the lack of knowledge of the electrical energy since we do
necessary,
to characterize some important parameters as Rin
not have biological sensors and so we have big difficulties to
and
the
maximum
input voltage and calibrate it, to verify its
estimate it [6].
transfer
function
and,
in case it is necessary, equalize it by
To compensate this situation the first thing to do is
software.
constantly measuring the electrical net evaluating the PQ
A preventive calibration of the instrumentation, necessary to
parameters. Currently this is difficult to obtain because of the
characterize
the audio card, it is also necessary. We used a
high costs of high accuracy PQ analyzers [7].
Yokogawa FG120[8] function generator calibrated, both in
I.

INTRODUCTION

amplitude and in frequency, in a specialized Anritsu laboratory

certified by the Italian Calibration System. For the first we
connected the function generator to a digital multimeter Fluke
8840A[9] with true RMS option and to a Lecroy Waverunner
LT342[10] oscilloscope to verify if the Yokogawa fixed VPP
sinusoidal waves output would be constant in the time. By this
test we verified that the function generator gives voltage linear
values. For the frequency calibration we used an Anritsu
MS2691A[11] spectrum analyzer. Thanks the flatness
function, it is possible to record and to show a frequency array
spaced of 50 Hz starting from 50 Hz up to 1500 Hz and pre
fixed on the function generator.
Figures 1a and 1b show the calibration bench and the linear
response of the Yokogawa on the oscilloscope monitor.

Yokogawa FG120

Lecroy LT342
a
b
Figure 1. Calibration instrumentations bench (a) and the response of the
Yokogawa FG120 function generator (b).

to: Rin = 11.8 mV/1.16 A = 10.17 k.

The procedure has been executed for both channels leaving
without load the channel not analyzed.
After that, to determine possible crosstalk phenomena
between the two input channels, we used a 1 k resistor on the
channel previously left idle, observing if on the other channel
was induced interference. Repeated this procedure for both
channels we analyzed the data by an FFT dont checking
interference.
Then we verified the input stage dynamic characterizing the
maximum input voltage to avoid saturation phenomena. We
fixed the Yokogawa sinusoidal output voltage to a frequency of
1 kHz and a incremental amplitude started from 0.2 VPP up to
the saturation limit characterized at 2.8 VPP. After this limit the
card saturates and distortion phenomena start to appear.
Another test allows us to determine the audio card low cut
off frequency: fixing the Yokogawa output voltage to a value
of 2 VPP we varied the frequency decreasing it starting from a
value of 1024 Hz up to 2 Hz. Fig. 3 shows that the low cut off
frequency is equal to 17 Hz (-3 dB) while to 50 Hz we have -1
dB. Fig. 3 shows the audio card low frequency response.
Frequency Hz

dB

Low pass frequency: 17 Hz

The function generator low secondary harmonics emissions

are reported in table 1 for a fundamental of 50 Hz:
TABLE I
SECONDARY HARMONICS EMISSION FOR YOKOGAWA FG120
Frequency
Amplitude

50 Hz
fundamental
0 dB reference
amplitude

100 Hz

500 Hz

1 kHz

1.5 kHz

Figure 3. Audio card frequency characteristic.

-73 dB

-80 dB

-82 dB

-85 dB

The latter test allows verifying the real sampling frequency

used by the audio card (nominally 8 ksamples/s). We used the
Yokogawa FG120 function generator to record in one second a
sinusoidal signal of 50 Hz, counting how many samples were
contained in ten periods. This procedure permits to determine
the exact sampling time to be correctly taken in account in
DFT.

After the function generator calibration we determined the

audio card characteristics that are summarized in table 2 for the
first PC.
TABLE II
AUDIO CARD CHARACTERISTICS
No distortion
input dynamic
Vmax = 2.8 VPP
[@ 1 kHz]

Nominal
input voltage

Input
Resistance

Low cut off

frequency

Crosstalk

Vn = 2 VPP

Rin =10 k

17 Hz

none

The first step has been to determine the input impedance of

the line-in building a measurement circuit by the following
electric scheme (Fig. 2).

B. The Analog Interface

We used, as voltage sensor, a transformer LEM LV 25-P[12]
while, as current sensor, a current transformer LA 55-P[13]
both suited for electronic measurements of DC and AC
voltages.

Current sensor
LA 55-P
USB
Figure 2. Measurement circuit to determine the audio card input impedance.

Applying a voltage reference of 1 V coming from the

function generator and using a measurement resistance of 5 k
(0.1%), we measured a voltage of 5.80 mV on the digital
multimeter. Using this it is possible to determine the current
that flows on the resistor equal to 1.16 mA and so the Rin equal

Stereo socket
for Audio Card
connection

Voltage sensor
LV 25-P

Domestic
electric
socket

Figure 4. Analogical interface circuit.

The first shows a bandwidth of 0 40 kHz while the second

0 200 kHz. They are active sensors and need a supply of 5

III.

V, carried out by a DC/DC integrated converter fed through the

USB port of the PC. Every sensor is connected to a channel of
the stereo socket of audio card by a stereo jack. In Fig. 4 is
shown the analog interface inserted in a proper box.

EARLY RESULTS

Current [A]

Voltage [V]

At the moment, two instruments are

active, placed in two different parts of
Roma: at Roma Tre University and in
an house in Torvaianica in the south
C. The Software
part of the city (respectively star and
~ 10 km
Exploiting the Windows DLL[15], in particular the square on the map of Fig. 6) where the
[Link], it is possible to manage the audio card to acquire PCs are normally used also as server.
data from the line-in. The software, developed in Visual Basic In the first case the load is the PC
6.0[16], allows us to choose the sampling frequency, to acquire itself, in the second case, the analogic
~ 30 km
data for a prefixed time window and to elaborate the data by a interface is connected to a multiple
DFT algorithm. To not overload the server and so to better socket connected to an A++ efficiency
menage the acquisition with the server activities of the PCs, we energy class refrigerant and an electric
decided to acquire 1 buffer every ten seconds. Each buffer is oven. There are no differences in the
composed by 8,192 samples obtained at a sampled frequency electrical energy supply, in fact both
of 8 kHz, frequency standard previewed by the acquire card, the buildings are simply connected to
Figure 6. Position of
instruments in Rome
for a total acquisition time equal to 1.024 s. After the the main without specific dedicated
territory.
acquisition the software elaborates the data and the next lines. We acquired data for 10 days
acquisition starts for 10 s after the previous. With this and the results of our analysis are following synthesized. In the
temporization the principal activities of the server is not limited. first case the RMS average voltage is 217 V with a standard
The software gives us the frequency of the fundamental and deviation of 1.095 V and it never exceed the limit. Being the
of the harmonics with a resolution of 0.1 Hz and with an PC a no linear load it hasnt significant to talk about power,
accuracy, joined to the previous audio card sample rate instead it is very interesting to see the current input due to its
calibration, equal to 20 ppm. It gives the true RMS value of PWM supplier. The next graph shows the voltage and current
Voltage and Current with accuracies of 0.2% and 0.5% trend for this load and it has been almost always the same in
respectively joined with the multimeter used during the these ten days.
Voltage
calibration. Moreover it gives the percentage of harmonics up
4
Current
to 24th respected the fundamental (THD) both for current and
300
voltage, Apparent, Active, Reactive and Deforming Power.
2
If the software doesnt identify an exceeding of norm
150
limits[14] for also only one of the analyzed parameters, the
samples are immediately deleted and the software records the
0
0
values of the parameters in a file with date and time.
-150
Otherwise, if the norm limits are exceeded, the software
-2
saves also a second file with the series of the samples acquired.
-300
In this way, it is possible to verify the samples to establish
-4
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
what problem has been occurred during the acquisition.
time [s]
During the program debugging, it has been useful to have a
Figure 7. Voltage and Current trend for a PC load.
visualization section of the program to check the exactness of
the data elaboration and to facilitate the visualization of the
How it can be seen, the voltage waveform is not a perfect
harmonics level. Fig. 5 shows the visualization section of the sine wave, but it is significantly distorted. This is clearly due to
program. This section is deactivated in the normal acquisition the kind of loads inside the University that are mainly no linear
process.
as computers and instrumentations.
1800

1600

Active Power

Reactive Power

Refrigerant activations

Oven use

1400

Active Power

1200

Power

Apparent Power

Spectrum
Analyzer
Monitor

Apparent Power

1000

800

600

Reactive
Power

400

200

Cos

15/02/09

Figure 5. Graphic section of the software.

16/02/09

Time (days)

17/02/09

18/02/09

Figure 8. Active, Reactive and Apparent trend for three days (15-18/02/2009)
of the second site.

For the second site the RMS average value is 238 V with a
standard deviation of 2.2 V and it exceeds the limit of 2.2%. In
this case the type of load makes extremely interesting the
power analyses. Fig. 8 shows the Active, Reactive and
Apparent Power of data acquired by the instrument during
three normal working days (15-18/02/2009).
It is possible to know the asynchronous activations and the
deactivations of the refrigerant and sometimes the activations
and the deactivations of the oven. In case of activation of the
only refrigerant the power factor has an average value of 0.78.
At the start of the activation, the apparent power reaches a peak
of 1.6 kVA. The contribution of the electric oven is almost
exclusively active power with a value of about 800 W and,
consequently, the power factor is practically equal to 1.
In both cases the frequency and the THD are below the norm
limits.
Table 3 shows, respectively for the two sites, how many
times, expresses in percentage respect the total number of
acquisitions, the nth harmonic exceeds the norms limit, the
average value respect the fundamental and the correspondent
standard deviation. As it is possible to see in the first case, the
level of the second harmonic often exceeds the limit. In the
second site, cause the high level of the fundamental, the
harmonics exceeds the limits few times than the first site.
TABLE III
HARMONIC DISTORTION TREND FOR UNIVERSITY (A) AND TORVAIANICA (B)
SITES
Harmonics
Order

Out of Limits %

Average % compared
the fundamental value

Standard
Deviation (V)

(A)

(B)

(A)

(B)

(A)

(B)

2nd

12.288

0.992

1.358

0.6993

0.601

0.363

3rd

0.004

1.584

0.4595

0.258

0.243

4th

0.004

0.036

0.222

0.2903

0.136

0.158

5th

1.104

1.5573

0.239

0.392

6th

0.013

0.4755

0.134

0.1618

0.072

0.103

7th

0.004

1.309

0.3813

0.220

0.178

8th

0.004

0.0011

0.111

0.1176

0.060

0.078

9th

0.477

0.946

0.2388

0.191

0.111

10th

0.004

0.109

0.0990

0.055

0.061

11th

0.945

0.1366

0.191

0.073

12th

0.009

0.085

0.0856

0.042

0.0449

13th

0.795

0.1093

0.207

0.060
0.042

th

14

0.004

0.065

0.0757

0.036

15th

0.319

0.0119

0.290

0.1611

0.095

0.088

16th

0.004

0.075

0.0668

0.040

0.039

17th

0.004

0.229

0.0926

0.085

0.050

th

18

0.004

0.059

0.0572

0.028

0.03

19th

0.138

0.0712

0.060

0.039

20th

0.004

0.048

0.0499

0.025

0.030

21rt

0.004

0.539

0.0687

0.028

0.036

22nd

0.004

0.049

0.0444

0.032

0.030

23rd

0.049

0.0575

0.025

0.032

24th

0.004

0.046

0.0408

0.022

0.026

How it could be expected, due to the non linear loads present,

the THD of the University site is higher than the second site.
The 2nd harmonic overshoots more often than the others
because the instrument site is close to a mechanic laboratory
running an old 1,5 kVA power drill that uses a direct current
brake for its AC motor. This brake uses a half-wave rectifier
that produces this type of distortion.
IV.

CONCLUSIONS

Both cause the increase of energy costs and a more high

sensibility towards energy problems, mainly linked with the
possible saving in case of anticipate breaking of electrical
devices , the needs to control and verify the quality of electrical
energy is more and more pressing.
In this article we presented a new instrument based on an
analogical self made interface connected between the mono
phase LV main and the PC audio card.
The instrument exploits the big calculus capacity of PC so
allowing an easy elaboration of the samples and an easy
memorization of the power quality normed parameters
analyzed. Exploiting an elaborative structure already existent
and bought for other applications, the instrument is extremely
cheap and so buyable also by LV, domestic or business,
customers. The low costs will make easy a rapid distribution on
territory of these instruments particularly recommended for
those little shops or offices that have a server PC continuously
active but not fully used. The instrument calibration process
has been also described.
Two instruments have been constructed and, at the moment,
they constantly acquire information about the net.
The early results show that a non linear load placed near the
first site produces an increase of the 2nd on the electrical LV
net, whose effects should be better analyzed and understood,
verifying them on all the other devices connected to the same
net. Moreover the voltage amplitude and current behaviors
over time justify our interests in a better knowledge of the
electric net by means of continuous measurements with the aim
to correlate possible electrical machine malfunctionings with
disturbances coming from the electrical net.
The next steps will be an improvement of the software that
has to include an automatic calibration section for the PC audio
card to avoid a tiring calibration job, and to increase the
acquisition points on territory to know as better as possible the
PQ parameters trend.
REFERENCES
[1] R.C. Dugan, M. F. McGranaghan, S. Santoso, H. W. Beaty, Electrical
Power Systems Quality, Second Edition, McGraw-Hill, 2004,vol. I,ISBN:
0-0713-8622-X.
[2] J. Arrillaga, N.R. Watson, S. Chen, Power System Quality Assessment,
John Wiley & Sons, 2000, ISBN 0 471 98865 0.
[3] E.F. Fuchs, M.A.S. Masoum, Power Quality in Power Systems and
Electrical Machines, ISBN 978-0-12-369536-9.
[4] M. Caciotta, F. Leccese, T. Trifir, From Power Quality To Perceived
Power Quality, The IASTED International Conference on Energy and

Power Systems EPS 2006, 29-31 March 2006, Chiang Mai, Thailand, pp.
94-102, ISBN CD: 0-88986-586-8.
[5] F. Leccese, Rome, a first example of Perceived Power Quality of
electrical energy: the telecommunication point of view, International
Telecommunications Energy Conference INTELEC 2007, September 30
October 4 2007, Rome, Italy, pp. 369-372, ISBN: 978-1-4244-1628-8,
IEEE Catalog Number: CFP07INTC, Library of Congress: 88-656128.
[6] F. Leccese, A first analysis of Perceived Power Quality for domestic
customers, 12th IMEKO TC1 & TC7 Joint Symposium on Man Science
& Measurement, September, 3 5, 2008, Annecy, France, ISBN: 29516453-8-4.
[7] Internet information about PQ Analyzers of several producers.
[8] Yokogawa FG120 User Manual.
[9] Fluke 8840A Digital MultimeterUser Manual.
[10] Lecroy Waverunner LT342 User Manual.
[11] Anritsu MS2691A Spectrum Analyzer User Manual.
[12] LEM LV 25-P Datasheet.
[13] LEM LA 55-P Datasheet.
[14] CEI EN 50160 Voltage Characteristics of Electricity supplied by public
distribution systems.
[15] MSDN: Microsoft Developer Network.
[16] Visual Basic 6.0 Manual.

Optimal Capacitor Placement on Interconnected

Distribution Systems in Presence of Nonlinear
Loads Using Genetic Algorithms
Ahmad Galal Sayed (1), Hosam K.M. Youssef (2), Member, IEEE
1

Engineering Consultants Group, Cairo, Egypt

Electric power and machine Dept., Faculty of Engineering, Cairo University, Egypt

Abstract- Due to simplicity of analysis of radial distribution

systems, all previous work studied the effect of nonlinear loads on
the optimal solution of the Capacitor Placement Problem (CPP)
on only radial distribution systems. The study of the optimal
capacitor placement on interconnected distribution systems in
presence of nonlinear loads using Genetic Algorithms (GA) is
presented in this paper. Results (power losses, operating
conditions and annual benefits) are compared with that obtained
from radial and loop distribution systems. Computational results
obtained showed that the harmonic component affects the optimal
capacitor placement in all system configurations. When all loads
were assumed to be linear, interconnected and Loop system
configurations offer lowest power losses and best operating
conditions rather than the radial system configuration while
radial system configuration offer best annual benefits due to
capacitor placement. In distorted networks, interconnected
systems configuration offer lower power losses, best operating
conditions and best annual benefits due to capacitor placement.

I. INTRODUCTION
Electric power is supplied to final users by means of
Medium Voltage (MV) or Low Voltage (LV) distribution
systems, their structures and schemes can differ significantly
according to loads location. Overhead lines with short
interconnection capabilities are mostly employed in rural areas,
whilst cables with a great number of lateral connections for
alternative supplies are widespread used in urban areas.
Most power distribution systems are designed to be radial, to
have only one path between each customer and the substation.
The power flows exclusively away from the substation and out
to the customer along a single path, which, if interrupted,
results in complete loss of power to the customer. Its
predominance is due to two overwhelming advantages: it is
much less costly than the other two alternatives (loop and
interconnected systems) and it is much simpler in planning,
design, and operation.
An alternative to purely radial feeder design is a loop
system, which has two paths between the power sources
(substations, service transformers) and each customer.
Equipment is sized and each loop is designed so that service
can be maintained regardless of where an open point might be
on the loop. Because of this requirement, whether operated

radially (with one open point in each loop) or with closed

loops, the basic equipment capacity requirements of the loop
feeder design do not change. In terms of complexity, a loop
feeder system is only slightly more complicated than a radial
system. Power usually flows out from both sides toward the
middle, and in all cases can take only one of two routes.
Voltage drop, sizing, and protection engineering are only
slightly more complicated than for radial systems.
Interconnected distribution systems are the most complicated
and costly but most reliable method of distributing electric
power. An interconnected distribution system involves multiple
paths between all points in the network. Interconnected
systems provide continuity of service (reliability) far beyond
that of radial and loop designs: if a failure occurs in one line,
power instantly and automatically re-routes itself through other
pathways.
Interconnected distribution systems are more expensive than
radial distribution systems, but not greatly so in dense urban
applications, where the load density is very high, where the
distribution must be placed underground, and where repairs
and maintenance are difficult because of traffic and congestion,
interconnected systems may cost little more than loop systems.
Interconnected systems require little more conductor capacity
than a loop system. The loop configuration required "double
capacity" everywhere to provide increased reliability.
Interconnected systems is generally no worse and often needs
considerably less capacity and cost, if it is built to a clever
design and its required capacity margins are minimized.
The solution procedures of the Capacitor Placement Problem
(CPP) start with performing a load flow analysis to analyze the
steady-state performance of the power system prior to capacitor
placement and after capacitor placement and to study the
effects of changes in capacitor sizes and locations.
The biggest advantages of the radial system configuration, in
addition to its lower cost, are the simplicity of analysis and
predictability of performance. Because there is only one path
between each customer and the substation, the direction of
power flow is absolutely certain. Equally important, the load
on any element of the system can be determined in the most
straightforward manner by simply adding up all the customer
loads "downstream" from that piece of equipment. Because

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
13

load and power flow direction are easy to establish in radial

distribution system, voltage profiles can be determined with a
good degree of accuracy without resorting to exotic calculation
methods; equipment capacity requirements can be ascertained
exactly; capacitors can be sized, located, and set using
relatively simple procedures (simple compared to those
required for similar applications to non-radial (loop and
interconnected) system designs, in which the power flow
direction is not a given).
The interconnected distribution systems are much more
complicated than other forms of distribution, and thus much
more difficult to analyze and operate. There is no
"downstream" side to each unit of equipment. This complicates
load estimation, power flow analysis, and protection planning.
Due to simplicity of analysis of radial distribution systems,
all previous work studied the effect of nonlinear loads on
optimal solution of CPP on only radial distribution systems.
The study of the optimal placement and sizing of fixed
capacitor banks placed on distorted interconnected distribution
systems using Genetic Algorithms (GA) is presented in this
paper. Results (power losses, operating conditions and annual
benefits) are compared with that obtained from radial and loop
distribution systems. The radial, loop and interconnected
distribution systems models are obtained by suitably
simplification of a typical Italian grid. Commercial package
ETAP PowerStation 4.0 program is used for harmonic load
flow analysis, the proposed solution methodology is
implemented in Microsoft Visual Basic 6 programming
language.
II. PROBLEM FORMULATION
A. Assumptions
The optimal capacitor placement problem has many
variables including the capacitor size, location, capacitor cost,
voltage and harmonic constraints on the system. There are
switchable capacitors and fixed type capacitors in practice.
However, considering all variables in a nonlinear fashion will
make the placement problem very complicated. In order to
simplify the analysis, only fixed type capacitors are considered
with the following assumptions:
1) The system is balanced.
2) All the loads are constant
B. Capacitor size and cost
Only the smallest standard size of capacitors and multiples
of this standard size are allowed to be placed at the buses to
have more realistic optimal solution. The capacitor sizes are
treated as discrete variable and the cost of the capacitor is not
linearity proportional to the capacitor size, this makes the
formulated problem a combinatorial one.
C. Objective Function
The objective of the capacitor placement problem is to
reduce the total energy losses of the system while striving to
minimize the cost of capacitors installed in the system. The
objective function consists of two terms. The first is the cost of

the capacitor placement and the second is the cost of the total
energy losses.
The cost associated with capacitor placement is composed of
a fixed installation cost, a purchase cost and operational cost
(maintenance and depreciation). The cost function described in
this way is a step-like function rather than a continuously
differentiable function since capacitors in practice are grouped
in banks of standard discrete capacities with cost not linear
proportional to the capacitor bank size.
It should be pointed that since the objective function is nondifferentiable, all nonlinear optimization techniques become
awkward to apply.
The second term in the objective function represents the total
cost of energy losses. This term is obtained by summing up the
annual real power losses for the system.
D. Operational Constraints
Voltages along the feeder are required to remain within
upper and lower limits after the addition of capacitors on the
feeder. Voltage constraints can be taken into account by
specifying the upper and lower bounds of the magnitude of the
voltages. The distortion of voltage is considering by specifying
for maximum total harmonic distortion (THD) of voltages and
the maximum number of banks to be installed in one location is
taken into account.
E. Mathematical Representation
The capacitor placement problem is expressed
mathematically as shown below:
J

Min

F = K p Ploss + K cj Q cj

(1)

j =1

Subject to:

Vmin V j Vmax
THD j THD max

Q Q
c
j

c
max

(2)
(3)
(4)

Where
c
Qmax
= L Q0c

Kp

Ploss
J

K cj
Q

(5)

The total annual cost function.

c
j

Vj

Annual cost per unit of power losses.

The total power losses.
(Result from ETAP PowerStation Harmonic Load
Flow Program).
Number of buses.
The capacitor annual cost/kvar.
The shunt capacitor size placed at bus j.
The rms voltage at bus j.
(Result from ETAP PowerStation Harmonic Load

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
14

Flow Program).

Vmin

Minimum permissible rms voltage.

Vmax

Maximum permissible rms voltage.

THD j

The total harmonic distortion at bus j.

(Result from ETAP PowerStation Harmonic Load
Flow Program).

THDmax Maximum permissible total harmonic distortion.

c
Qmax
Maximum permissible capacitor size.
L
Q0c

An integer.
Smallest capacitor size.

Bounds for (2), (3) are specified by the IEEE-519 standard

[1].
III. GENETIC ALGORITHM
A. Solution Algorithm
Combinatorial optimization problems can be solved either
by exact or by approximate methods. In exact methods, all the
feasible solutions are evaluated and the best one is selected as
the optimal solution. However, exact methods are impractical
when a real-life problem is to be evaluated. In this paper, GA
will be used for the solution of CPP.
B. Genetic Algorithm Framework
There are four components in the design of a GA-based
solution methodology. These include the initialization of the
algorithm, fitness evaluation, selection and genetic operators.
Algorithm initialization is the process of randomly
generating a set of initial feasible solutions forming the socalled "initial population". The number of these solutions is
referred to as the "population size". Each iteration in a genetic
algorithm, known as a "generation", results in a new set of
feasible solutions.
Genetic algorithm needs some fitness measure to determine
the relative 'goodness' of a particular solution. This can be
obtained either by direct evaluation of the objective function or
by some other indirect means. Fitness evaluation is the
criterion guiding the search process of a genetic algorithm.
In genetic algorithms, parents are selected to produce
offspring. Selection process can be carried out in different
ways as discussed before in chapter three.
Genetic operators are the probabilistic transition rules
employed by a genetic algorithm. A new and improved
population is generated from an old one by applying genetic
operators. Operators used by genetic algorithms include
crossover and mutation.
Crossover is the process of choosing a random position in
the solution and swapping the characters around this position

with another similarly partitioned solution. The random

position is referred to as "the crossover point". In other words,
crossover defines the outcome as gene exchange. Crossover
operator proved very powerful in genetic algorithms. Mutation
is the process of random modification of a particular value of a
solution with a small probability. Mutation is applied to alter
some genes in the solutions. When a gene exchange resulting
from application of a crossover operator is not meeting
appropriate restriction, mutation might be very helpful in
providing a proper gene exchange amendment. Mutation is
generally seen as a background operator that provides a small
amount of random search. It increases the population diversity.
It also helps expand the search space by reintroducing
information lost due to premature convergence. Therefore, it
drives the search into unexplored regions.
In addition to the above components, the stopping criterion
of the algorithm is of great significance. It determines when the
algorithm shall be stopped or terminated and thus, considering
the best solution obtained so far as the optimal solution.
C.

Design of a successful GA-based solution methodology

In designing a GA-based solution methodology, several
decisions concerning the algorithm parameters shall be
properly made in order to obtain high-quality solutions.
Premature convergence to local optimum may result if the
algorithm parameters are not selected in an appropriate
manner.
Population size and the way the initial population is selected
will have a significant impact on the results. Initial population
could be seeded with heuristically chosen solutions or at
random. In either case, the initial population should contain a
wide variety of structures. The population size and the initial
population are selected such that the solution domain
associated with the population is sufficiently covered. The
population size depends on the criteria for selecting the initial
solutions. A constant size population of solutions shall be
judged by the algorithm designer. If the population size is too
small, the solution domain will not be adequately searched and,
thus, resulting in poor performance. Premature convergence to
local solutions can be prevented by using a large population
size. However, this may slow down the convergence rate.
The performance of a genetic algorithm is highly sensitive to
the fitness values. The fitness value may be selected as the
objective function to be optimized. Some researchers, however,
believe that the objective function value is a naive fitness
measure. Therefore, using the objective function value
associated with each solution as a fitness measure is rarely a
good idea to some researchers. When applying a crossover
operator, the resulting offspring can be either feasible of
infeasible. There are two ways to deal with infeasible solutions.
One way is to design heuristic operators transforming
infeasible solutions to feasible ones. The second is by
penalizing this infeasibility in the objective function. In the
latter case, selection of proper penalty factors is another
decision to be made by the algorithm designer.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
15

Crossover rate and mutation rate play a very important role

in the performance of genetic algorithm. A higher crossover
rate introduces new solutions more quickly into the population.
If the crossover rate is too high, high-performance solutions are
eliminated faster than selection can produce improvements. A
low crossover rate may cause stagnation due to the lower
exploration rate. Mutation rate shall not be too high in order
not to prevent crossover from doing its work properly. Some
researchers reported that a variable mutation rate rather than a
fixed one is more beneficial to be utilized. In the initial stage of
the GA, it is recommended to start with a high crossover rate
and a low mutation rate since the crossover is mainly
responsible for the search at this stage. As the algorithm
progresses, the crossover operator becomes less productive and
therefore, the mutation rate shall be increased.
Simple GA involving just mutation and crossover has proved
to be quite powerful enough. Some designers proposed to add
an inversion operator where a section of a solution is 'cut out'
and then re-inserted reversely. However, it has not been found
significantly useful. It may be used instead of mutation
operator to explore new regions of the searching domain.
As pointed out earlier, newly generated solutions replace
existing solutions in subsequent generations of GA. Two
replacement approaches are available to the algorithm designer
to select among which. These are the incremental or 'steadystate' approach and the generational approach. In incremental
or steady state replacement, once a new feasible solution has
been generated, it will replace an above-average fitness value
member chosen randomly. Members that fitness values are
better than the calculated average are transferred to the next
generation with no change. In generational replacement a new
population of children is generated to replace the whole parent
population. The steady-state replacement approach has the
following advantages.
1. If steady-state replacement approach is not employed,
there will be no guarantee that the best solution in the
current population will survive into the next generation.
With this approach, best solutions are always kept in the
population and the newly generated solution is
immediately available for selection and reproduction.
2. It is more efficient when compared to the generational
approach. Faster convergence is usually expected with this
approach.
3. It prevents the occurrence of duplicates. Duplication is
unhelpful since it wastes resources on evaluating, the same
fitness function and it distorts the selection process by
giving extra chances to the duplicate solutions to
reproduce.
Researchers, however, reported some success with the
generational replacement approach. The genetic algorithm can
be designed to stop if a pre-specified number of iterations are
completed or if no improvement is encountered in the optimal
solution during a given number of consecutive iterations. The
stopping criterion shall be properly tuned with the other
parameters like the population size, crossover rate and

mutation rate in order to obtain a high quality solution.

D. Application of GA to the CPP in distorted distribution systems
Because of its simplicity, generality and ability to cope with
practical constraints, a genetic algorithm has been designed to
solve the general CPP in a distribution system. The following
remarks shed some light on the design aspects of the algorithm
as applied to the CPP:
The population size is a fixed value and the same is
determined empirically by trial and error process.
The objective function itself is used to provide fitness
values of the newly generated solutions. Once a new
solution is generated, its associated feasibility is checked.
If the solution is infeasible, the penalty factor is applied.
Selection, crossover and mutation are applied as genetic
operators in the algorithm design. No additional operators
such as the inversion operator are considered in the design
process.
The algorithm is designed based on a fixed, rather than a
variable, mutation rate throughout the search.
Steady-state replacement approach is selected in the design
of the algorithm.
The algorithm is designed to stop after predetermined
number of generation.
Based on the above remarks, a GA-based solution
methodology applied to the CPP has been implemented. The
algorithm implementation can be summarized as follows:
In this paper the representation by means of strings of
integers was chosen. Each gene(i) (represent buses) of the
chromosome (its length is equal to the total number of the
system buses (m)) can store a 0, which indicates absence of
capacitors on the corresponding bus or an integer different
from 0 that indicates the number of added capacitor sizes that
is added in the bus(i). Therefore a chromosome can be
represented as Figure (1)
Bus number
1
2
3
...
...
m-2
m-1
m
gene
0
3
4
...
...
2
1
3
Figure 1: A genetic algorithm chromosome of the capacitor placement problem

The algorithm procedure can be summarized as follow:

1. Input GA parameters, i.e. population size, crossover
rate, mutation rate, selection type, crossover type,
mutation type and termination mode.
2. Generate a set of initial solutions, forming the initial
population, randomly.
3. Run ETAP Power Station 4.0 program to calculate
system power losses, bus voltages and total harmonic
distortion at each bus.
4. Calculate the associated fitness value (objective
function) of each solution.
5. Check the constraints and applying the penalty factor
(by adding 1E10 to the fitness value) if the constraints is
not satisfied.
6. Create a new population by performing selection,

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
16

7.

crossover and mutation on the individuals.

Discard the old population and iterate using the new
population if the stopping criterion is not satisfied.
IV. NUMERICAL RESULTS

source typical IEEE- XFMR Magnet.

Kp was selected to be 168 $/kW, and the voltage limits on
the rms voltages were selected as Vmin= 0.95 pu, and
Vmax=1.05 for case (1) and Vmin= 0.93 pu, and Vmax=1.05
for case (2). The maximum THD was selected as THDmax = 5
B. Test System results

A. Test System data

The distribution network models are obtained by suitably
simplification of a typical Italian grid [2]. Single line diagram
of the network is shown in Fig. (2) and the system data as
follows:
Two 132 kV HV networks with the same short circuit power
MVAsc of 6000 MVA;
Two HV/MV substations, comprising each a 132 kV HV
busbar, a 132/20 kV 40 MVA transformer and a 20 kV MV
busbar;
A feeder, subdivided in three line sections (L01, L12 and
L23) of 3 km each with % positive sequence impedance (100
MVA base) R=5.17, X=4.23, Z=6.68
A series of further passive overhead feeders;
Link lines between various feeder (Lm1 and Lm2);
Configuration switches (S1, S2, S3, S4, S5 and S6).
Table (1) show the system load data
The proposed GA was applied to the test system-2 for three
different network configurations:
1. Radial configuration (S1 open, S2 open, S3 open , S4
open , S5 open and S6 open);
2. Loop configuration (S1 open, S2 open, S3 close , S4
close , S5 close and S6 close);
3. Interconnected configuration (S1 close, S2 close, S3
open , S4 open , S5 open and S6 open).
Table (1): Test system load data
1
2
3
4
5
6

Load (80% motor, 20% static)

MW
MVAR
5.25
5.356
5.25
5.356
5.25
5.356
4.5
2.18
4.5
2.18
4.5
2.18

7
8
9

4.5
4.5
4.5

2.18
2.18
2.18

10
11

4.5
4.5

2.18
2.18

12

4.5

2.18

Bus No.

Commercially-available capacitor sizes with real costs/kvar

were used in the analysis. It was decided that the largest
capacitor size Qcmax should not exceed the total reactive load,
i.e., 35688 kvar. The yearly costs of capacitor sizes as
described in [3]
Optimum shunt capacitor sizes have been evaluated for the
following cases:
Case (1): All loads are assumed to be linear
Case (2): Each 40 MVA transformer has harmonic current

Table (2) Comparison of results between radial, loop and

interconnected distribution systems for case (1)
Variable

Before OCP
Radial

Loop

QC1 (kvar)
QC2 (kvar)

-----

QC3 (kvar)
QC4 (kvar)
QC5 (kvar)

Case (1)
Radial

Loop

-----

Intercon
nected
-----

4050
3900

3600
4050

Intercon
nected
3900
4050

-------

-------

-------

4050
1500
3300

3900
1800

3600
3600

2400

3300

QC6 (kvar)
QC7 (kvar)

-----

-----

-----

1650
3450

2700
3300

2100
3450

QC8 (kvar)
QC9 (kvar)

-----

-----

-----

750
3450

3300

2400

3450

2400

QC10 (kvar)

---

---

---

1800

QC11 (kvar)
QC12 (kvar)

-----

-----

-----

2250
2700

2400
2550

1200
2100

1650

2250

Total
capacitor
(kvar)
Min.
voltage(pu)
Max.
voltage(pu)
Power losses
(kW)
Cap. cost
($/year)
Total cost
($/year )
Benefits ($/
year )

---

---

---

32850

35100

34350

0.91

0.91484

0.92671

0.96535

0.97364

0.97521

0.9551
6
1244.1

0.95516

0.94462

0.98778

0.98860

0.98554

1211.1

1206.5

812.8

795.7

796.8

---

---

---

6142.8

6301.65

6156.9

20900
8
---

203464

202692

142685

139904

139952

---

---

66323

63560

62739

Figure 2: Test system - Simplified Typical Italian Grid

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
17

Table (3) Comparison of results between radial, loop and

interconnected distribution systems for case (2)

40000
35000
30000

Before OCP

Variable
Radial
QC1 (kvar)
QC2 (kvar)

-----

-----

QC3 (kvar)
QC4 (kvar)
QC5 (kvar)

-------

-------

-------

2550
1350
2100

3450
2100

3150
2700

1650

1800

QC6 (kvar)
QC7 (kvar)

-----

-----

-----

2100
300

2850
2400

750
1650

QC8 (kvar)
QC9 (kvar)

-----

-----

-----

2850
2550

2100

1800

2550

1650

QC10 (kvar)

---

---

---

3450

QC11 (kvar)
QC12 (kvar)

-----

-----

-----

900
3000

2400
1350

3000
3600

2400

300

Total capacitor
(kvar)
Min.
voltage(pu)

---

---

---

26850

27300

27450

0.9099

0.9148

0.92669

0.948

0.9544

0.9649

0.9551

0.9551

0.94461

0.98657

0.9866

0.9772

2.68
1244.2

2.68
1211.1

2.65
1206.5

4.95

4.99

4.96

873.7

844.928

828.92

---

---

---

5007

4992

5135

20903
1
---

20347
4
---

202707

151788

146940

144413

57242.9

56534

58293

---

Radial

Loop

2700
3000

900
3150

20000

Intercon
nected
-----

Max.
voltage(pu)
Max. THD (%)
Power losses
(kW)
Cap. cost
($/year)
Total cost
($/year )
Benefits ($/
year )

Loop

Intercon
nected
3600

860
840
820

Power losses (KW/Year) Case(1)

800

Power losses (KW/Year) Case(2)

780
760

10000
5000
0

Radial

Loop

The study of the optimal capacitor placement on

interconnected distribution systems in the presence of
nonlinear loads using Genetic Algorithms (GA) is presented in
this paper. Results (power losses, operating conditions and
annual benefits) are compared with that obtained from radial
and loop networks. The radial, loop and interconnected
distribution systems models are obtained by suitably
simplification of a typical Italian grid. Computational results
obtained showed that the harmonic component affects the
optimal capacitor placement in all system configurations.
When all loads were assumed to be linear, interconnected and
Loop system configurations offer lowest power losses and best
operating conditions rather than the radial system configuration
while radial system configuration offer best annual benefits
due to capacitor placement. In distorted networks,
interconnected systems configuration offer lower power losses,
best operating conditions and best annual benefits due to
capacitor placement.
REFERENCES
[1]

[3]

68000
66000
64000
62000
60000

Benefits($/year) - Case(1)

58000

Benefits($/year) - Case(2)

Interconnected

V. CONCLUSION

Interconnected

Figure 3: Power losses (Kw/year) for case (1) & (2)

Loop

Figure 5: Total sizes of installed capacitors (Kvar) for case

(1) & (2)

[2]

740

Total sizes of installed

capacitors (Kvar)- Cas e(2)

15000

3450

880

Radial

Total sizes of installed

capacitors (Kvar) - Case(1)

25000

Case (2)

IEEE Recommended Practices and Requirements for Harmonic Control

in Electrical Power Systems, IEEE Std. 519-1992, 1993.
[Link], [Link] and [Link], Non-conventional Distribution
Network Schemes Analysis with Distributed Generation, TEQREP,
Bucharest, 2004
Y. Baghzouz, S. Ertem, Shunt Capacitor Sizing for Radial Distribution
Feeders with Distorted Substation Voltages, IEEE Trans. on Power
Delivery, Vol. 5, No. 2, April 1990, pp. 650-657.

Ahmad Galal Sayed received the [Link]. degree in electrical power and
machines from the Faculty of Engineering at Alexandria University, Egypt, in
2003. He is currently working toward the [Link]. degree in the department of
electrical power and machines, Faculty of Engineering, Cairo University. His
research interest includes computer aided analysis of power systems, and
power system harmonics.

56000
54000
52000
50000
Radial

Loop

Interconnected

Figure 4: Benefits ($/year) for case (1) & (2)

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
18

Computer Relaying Software Reliability Evaluation

Using Event Tree
Aditya Anil#, D.K Mohanta#, D.S Roy*1, A.K Panda*2
#

Department of Electrical and Electronics Engineering

Birla Institute of Technology (Deemed University)
Mesra, Ranchi-835215, India
*1Department of Computer Science and Engineering
*2Department of Electronics and Communication Engineering
National Institute of Science and Technology,
Berhampur - 761008, India

Abstract- Software reliability evaluation has emerged as an

area of importance in recent times with the proliferation of
software based systems. But highly dependable software systems
used for safety critical application, such as computer relays for
power system transmission line protection produce little failure
data. Hence a statistical method known as statistics of extreme is
used for reliability evaluation. This paper deals with calculation of
exceedance probability of computer relay and using this data an
event tree of computer relay has been drawn.

I.

INTRODUCTION

The reliability evaluation of software based systems has

become a major concern. Hardware reliability can be resolved
much consistently compared to software reliability, which is a
rather less explored field until recent times [1]. The reliability
assessment of software requires the previous history of failure
data based on which a statistical estimation is put forward.
Nevertheless for novel and highly dependable software
systems, such failure data is scarcely available and thus
formulation of statistical estimation of reliability index from
such meager available data does not seem viable [2].
Statistics of the extremes is a special branch of statistics
which deals with scenarios involving extreme events and can
model distribution of rare events, those for which very few or
no available data is present to work with as in the case of
computer relays software evaluation where the failure data is
very less.
For determining the design, procedural weakness & various
hazardous consequences due to faults (initiating accidental
event) event tree is developed for computer relays software.
An event tree is an inductive procedure that shows all the
possible outcomes, resulting from an accidental or initiating
event, taking into account whether all the safety barrier are
functioning or not.
In this paper, the software reliability evaluation of a computer
relay is investigated, whose algorithm is based on the use of
wavelet-fuzzy combined approach for detection, classification
and location of faults on a transmission line. With the
extensive use of fast and accurate digital signal processing
techniques such as the wavelet transform, detection and

classification of different faults have been done very easily

[3][4].
II. SOFTWARE RELIABILITY EVALUATION
Computers have pervaded to all aspects of modern society.
Practically speaking, the existence of software based systems
in our lives is omnipresent and the size and complexity of
computer-intensive systems are growing day-by-day. Thus the
reliability evaluation of software based systems has become a
major concern. Especially for safety critical system like
computer relay, reliability evaluation is extremely necessary.
But as the number of fault found in the different modules of
software is very rare, hence for reliability evaluation statistics
of extreme is used.
In this paper, the software reliability evaluation of a computer
relay has been obtained, the computer relay has three modules
and Monte Carlo simulation was employed for incorporating
the stochastic nature of fault occurrence in the system. Using
this data, the failure probability and the corresponding Cdf for
each of the three module of computer relay were calculated.
By using the principles of statistics of extreme failure
probability and exceedance probability were found out.
A. Statistics of Extreme
Software reliability modeling is a long-established approach,
wherein a systems past failure data is analyzed for prediction
of future behavior [5]. But software systems that are supposed
to be built for highly dependable applications is difficult to
analyze using conventional software reliability growth models
owing to reasons such as the novelty of the software to be
developed, sparingly few available failure data.
Statistics of extremes offers a modeling scheme that is
independent of the test sampling order, the amount of available
data and furthermore it does not require a priori knowledge
about the underlying distribution of any parameter. a
distribution F converges to some asymptotic form in its
extreme tails [5]. Condition that a distribution F converges to
an asymptotic form in its maximum tail if

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
19

(1)

Lim FXn ( an + bn x) = H ( x)
n

For applying statistics of extremes, the first step is to plot the

empirical cumulative distribution function (Cdf) of available
failure data in a Gumbel type probability paper [6] to find out
its domain of attraction. Reference [7] shows that for most
distributions the domain to attraction belongs to any one of the
three Gumbel type asymptotic families, namely the Gumbel
Type I, Gumbel Type II or Gumbel Type III [6]. The next step
is to determine the software reliability estimate for highly
dependable system employing a graphical based analysis using
the form of empirical Cdf plot.
For obtaining the reliability indices needed for software
evaluation, failure probability and exceedance probability is
needed. These indices are calculated using the following
equations [8]:
(2)
FS ( s ) = exp ( exp ( s ) )
x
Where s = k ln
= ln[ ln[ FS ( s )]]
w
s
K=
ln( x ) ( )
FS ( s ) = F

X1

( X 1 x) = 1 exp[(

(3)
(4)
(5)

P X 1 ( X 1 x) = [1 F X 1 ( X 1 x) ]
(6)
Where is the characteristic smallest variable of initial variate
x, is the lower bound which is assumed to be 1 in this case
and k is the inverse measure of dispersion of initial variate x.
F X 1 ( X 1 x) is the probability that the number of failure is

III.

Figure 1. Gumbell Type III (Empirical Cdf) plot for Detection Module
TABLE 1 Cdf data of Detection Module
Fault type
Fault #
Probability
AG (LI )

x K
) ]

less than x (failure probability). Whereas P

the exceedance probability of the software.

TABLE I

X1

( X 1 x) gives

CASE STUDY AND RESULT

Fault data is obtained from the Monte Carlo simulation result,

from this fault data corresponding Cdf has been calculated for
each of the three modules of the computer relay. Here detail
calculations of detection module have been shown.
From the Cdf plot shown in Fig.(1) characteristic smallest
value has been calculated which comes out to be 38 for
detection module. Therefore (x-) is 37. The value of S can be
extrapolated using Fig. (1). The probability of having less
than (x - ) = 24 software failure is approximately = 0.80. i.e.,
Fs(s) = 0.80 as shown in table 1. Calculating S for (x - ) = 24
gives S = 1.5. K is calculated using eqn. (4), which comes out
to be 3.47. Therefore the failure probability expression for
detection
module
is
given
as
x 3.47
F X 1 ( X 1 x) = 1 exp
from this expression by
37
putting different values of x; failure probability for different
potential software failure for detection module can be obtained
and using eqn. (6) exceedance probability is calculated for
different number of software failures and tabulated in table 2.

.033

CDF
.0011

AG (MI)

.066

.003

AG (HI )

0.1

.0066

BG( LI)

0.133

.0111

BG(MI)

0.166

.0166

BG (HI)

0.2

.0233

CG (LI)

.233

.0310

CG (MI)

.266

.0399

CG (HI)

.300

.0499

ABG (LI)

10

.333

.0610

ABG (MI)

11

.366

.0732

ABG (HI)

12

.40

.0865

BCG (LI)

13

.433

.1010

BCG (MI)

14

.466

.1165

BCG (HI)

15

.500

.1332

CAG (LI)

16

.533

.1509

CAG (MI)

17

.566

.1698

CAG (HI)

18

.600

.1898

AB (LI)

19

.633

.2109

AB (MI)

20

.666

.2331

AB (HI

21

0.7

.2564

BC (LI)

22

0.733

.2809

BC (MI)

23

0.766

.3064

BC (HI)

24

0.80

.3331

CA (LI)

25

0.833

.3608

CA (MI)

26

0.866

.3897

CA (HI)

27

0.90

.4197

ABC (LI)

28

0.933

.4508

ABC (MI)

29

0.966

.4830

ABC (HI)

30

.5163

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
20

Table [Link] Probability of Detection Module

Exceedance
Z : No. of Potential lifetime software failure
Probability

Px ( X

x)

12

15

0.99999

0.9996

0.9951

0.9853

0.9663

For fault classification modules, from the Cdf plot in fig. 2,

the smallest characteristic value, comes out to be 31 & (x - )
=30. As in the previous calculation for detection module, from
the Cdf data, and the Cdf plot shown in Fig. 2 a value of S is
extrapolated. The probability of having less than (x - ) = 32
software failure is approximately 0.3299. Calculating S for (x ) = 32 gives S = -0.1034. Thereafter from eqn. (4), K is found
out to be 1.603. Thus, the analytical expression for the failure
probability
for
classification
module
1.603
x
is F X 1 ( X 1 x) = 1 exp

30
Using the eqn.(6) exceedance probability values are tabulated
in table 3.

In Fig. 3 the Cdf plot for fault location module is shown.

The extrapolated value of S is 0.66 and k comes out to be
1.741. Therefore the final expression obtained for failure
probability for fault location modules comes out to be
x 1.74
F X 1 ( X 1 x) = 1 exp
. Using eqn. (6) the
33
exceedance probability for fault location is tabulated in table 4.

Figure 3. Gumbell Type III (Empirical Cdf) plot for fault location Module

Table [Link] Probability of fault location Module

Exceedance
Z : No. of Potential lifetime software failure
Probability

Px ( X

x)

15

23

31

0.9977

0.9349

0.7986

0.6103

0.4286

Figure 2. Gumbell Type III (Empirical Cdf) plot for classification Module

Table [Link] Probability of classification Module

Exceedance
Z : No. of Potential lifetime software failure
Probability

Px ( X

x)

2
0.9957

8
0.9158

14
0.7808

27
0.4550

36
0.2743

For software reliability evaluation of computer relays

software, event tree is used, as shown in fig (4). Event tree
gives us the insight knowledge of the consequences, due to
initiating fault [9]. These consequences will depend upon the
functioning of the safety barriers. Here in this case for a
computer relaying, its three modules namely detection,
classification and fault location are the safety barriers.
Therefore, if all the three safety barriers are working, then the
system will be considered as a Risk free or healthy system.
The main job of a relay is to detect the faults. Hence if the first

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
21

safety barrier, which is detection module of a computer relay,

fails then the consequence will be termed as Risky, even if the
other two modules are operational. And if first safety barrier is
working and other two or one of the two is non-functional then
the system is in marginal state.
From the event tree, the probability that the accidental event
(fault) will lead to unwanted consequences if the safetybarriers are non functional is computed. Also probability of
Risk free state of computer relay is calculated. These values
are tabulated in table 5.
For getting these probability values the exceedance
probability for two numbers of failures has been taken for each
of the three modules from table 2, 3 & 4 respectively.
Therefore when all the modules are in Up state the reliability
comes out to be 0.9935 and the relay is said to be in Risk Free
state & when all are in down state reliability is 9.89*10-11 .

Table 5 State Probability of Event Tree

State Fault
Fault
Fault Probability
Detection Classification location
1

Down

Down

Down

Down

Up

4.29*10-8

Down

Up

Down

2.29*10-8

Down

Up

Up

9.93*10-6

Up

Down

Down

9.88*10-6

Up

Down

Up

4.29*10-3

Up

Up

Down

2.29*10-3

Up

Up

Up

0.9935

IV.

Down 9.89*10-11

CONCLUSION

The reliability evaluation for computer relaying demonstrates

the efficacy of the proposed approach for assessing the risk
associated with such software which rarely fails, but whose
failures lead to catastrophic consequences such as blackouts.
By using event tree, various state probabilities are evaluated
for system planning. Also the least risk state probability serves
as an indicator regarding the system reliability involving safety
critical applications such as computer relay.
REFERENCES
[1] D. S. Roy, D. K. Mohanta, A. K. Panda, Software reliability allocation
of digital relay for transmission line protection using a combined system
hierarchy and fault tree approach, IET Software, October, 2008, vol. 2, no. 5,
pp. 437 445.
[2] E. J Gumbel., Statistics of extremes, Columbia Press, New York, 1958
[3] M Jayabharata Reddy., D.K Mohanta.: A wavelet-neuro-fuzzy combined
approach for digital relaying of transmission line faults, Int. J. Electrical.
Power Component. System, 2007, 35, (12), pp. 13851407
Figure 4 Event Tree of Computer Relay

[4] M Jayabharata Reddy. D.K Mohanta., Performance evaluation of adaptive

network based fuzzy inference system approach for location of faults on
transmission lines using Monte Carlo simulation, IEEE Trans. Fuzzy System.,
2008, 16, (4)
[5] Michael R. Lyu (Ed.), Handbook of Software Reliability Engineering,
McGraw-Hill, 1995.
[6] E. Castillo, E. Alvarez, A. Cobo, T. Herrero, An Expert System for the
Analysis of Extreme Value Problems, 1993; University of Cantabria, Spain
[7] E. Castillo, Extreme Value Theory in Engineering, 1988; Academic Press
[8] L.M. Kaufman, J.B. Dugan, B.W. Johnson, Using statistics of the
extremes for software reliability analysis of safety critical systems, Proc.
Ninth Intl Symposium on Software Reliability Engineering, 1998 Nov, pp 355

- 363.
Marvin Rausand, System Reliability Theory, second edition, 2004, Wiley
& sons.

[9]

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
22

Relationship between Production Capacity of Wind

Power Stations and Needs of Ancillary Services
Vaclav Cerny, Andrea Fialova, Lucie Houdova, Petr Janecek and Eduard Janecek
Department of Cybernetics, Faculty of Applied Sciences, University of West Bohemia
Univerzitni 8, 306 14 Plzen, Czech Republic

AbstractNumber of wind power stations has been increasing

in the world for last years. Of course, this causes that electrical
power amount produced by this kind of power station and
subsequently delivered to electricity transmission systems has
also been growing. On one hand, this brings positive effects with
respect to living environment because production of coal, gas or
nuclear power stations can temporarily be reduced while wind
blows. On the other hand, this sets enhanced demands on needs
of ancillary services. This paper deals with these connections and
demonstrates obvious consequences that arise from them.

I. I NTRODUCTION
Wind energy produced by wind power stations is one of
renewable resources that has become very popular and used
world wide in recent years. It brings many benefits to mankind
particularly with respect to living environment. For instance,
production of other types of power stations such as coal, gas
or nuclear is supposed to be reduced due to wind blowing
because, as a result, wind power stations produce extra electrical power that is delivered to electricity transmission systems.
This implies that living environment is less polluted because
production of messy power stations is restrained.
Nevertheless, the problem is that time periods within which
wind is going to blow together with volume of blowing is
hardly ever reliably predictable to the future. This causes
that production of wind power stations cannot be scheduled
in advance. Therefore, the production is not usually taken
into account while electricity transmission system operation
planning performed by TSOs (Transmission System Operators) is under construction. On the contrary, extra electrical
power produced by wind power stations and delivered to
electricity transmission systems (while wind blows, of course)
is classified as a failure that exhibits by way of temporary
power surplus in the systems. Consequently, such a failure is
assumed to be decayed by employment of appropriate types
of ancillary services performed by TSOs as well. It is obvious
that this matter of fact puts increased requirements towards
needs of particular types of ancillary services.
Seeing that wind blowing strategy can be considered
as a random variable, the approach presented in the paper
and dealing with relationship between production capacity
of wind power stations and needs of ancillary services is
based on probability theory and mathematical statistics [1], [2].
Fundamentals of the approach focused on stochastic modelling
[3], [4] of electricity transmission system operation from both
technical and economical points of view have already been

established in other papers [5], [6], [7], [8]. Therefore, only

main ideas of the approach will be introduced at the beginning
of the paper. Subsequently, it will be presented that extra
electrical power delivered to electricity transmission systems
by production of wind power stations under the circumstance
that wind blows causes growth of power reserves of certain
types of ancillary services. In other words, TSOs are forced
to assure larger amount of certain types of ancillary services
when electricity transmission system operation planning is
under construction. It is obvious that this implies financial
cost increase with respect to operation of the system.
II. M AIN I DEAS OF THE A PPROACH
At the beginning, relevant monitored data and measured signals have been recognized from the system operation reliability
point of view. Since the reliability of the system operation is
particularly linked with power balance control, those data and
signals have included a power balance deviation in the closed
loop mode dPct (t), power control injections R(t) supplied to
the system by activations/deactivations of ancillary services,
starts and ends of forced power unit outages and time durations
related to the outages (times between failures and of repairs).
The deviation dPct (t) is then decomposed.
A. Decomposition of dPct (t)
It holds that:
dPot (t) = dPct (t) + R(t),

(1)

dPot (t) is a power balance deviation in the open loop mode.

The deviation represents behavior of the system without any
control actions and it is easily obtained from (1). The deviation
can be expressed as follows:
dPot (t) = dPo (t) + T (t),

(2)

dPo (t) is a power balance deviation in the open loop mode

of undisturbed system operation (without any forced power
unit outages) and T (t) constitutes the outages. The deviation
is obtained by excluding sectors where the outages occur
from dPot (t). Those sectors are identified in virtue of the
information about the starts and ends of the outages. The
components dPo (t) and T (t) are modelled.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
23

B. Probabilistic model of dPo (t)

Whereas dPo (t) is a random variable, it is modelled by
a proper distribution from a probabilistic viewpoint. It has
been proved by fit tests that the most proper distribution for
modelling dPo (t) is the Gaussian distribution:

1
fG (dPo ) =
e
2dPo (t)

[dPo dP (t)]2
o
2 2
(t)
dPo

(3)

fG (dPo ) is a probability density function. The parameters of

the distribution (mean dPo (t) and standard deviation dPo (t))
are computed from the real signal dPo (t).
C. Probabilistic model of T (t)
The forced power unit outages T (t) are modelled by the
Markov process with two states (see Fig. 1) [9], [10].

Fig. 1.

State diagram of Markov process

The transient rates are given by the relations:

1
1
, =
,
(4)
MTTF
MTR
M T T F , M T R are mean times between failures and of
repairs. The mean times are obtained through statistical processing of given times between failures and of repairs for
particular units. The model is described by the equations:

d
ps (t)

ps (t)
=
, ps (t) + pr (t) = 1,

pr (t)
dt pr (t)
(5)
ps (t) is a service probability and pr (t) a repair probability.
The forced outages are mostly caused by mechanical failures (e.g. seizure, leakage, fatigue, etc.). Therefore, dependencies among co-temporary outages are not considered in
the model. Partial outages are not taken into account either
because they are not very frequent, their durations are usually
short and amplitudes small. This means that they do not have
any substantial effect on the reliability of the system. Modelled
power units inhere in the bottom of the well - known bathtub
curve with respect to their life cycles. This means that the
failure and repair rates are about constant. Thus, the Markov
process is precise enough for the modelling.
=

D. Probabilistic model of dPot (t) and R(t) determination

Consequently, the probabilistic models of dPo (t) and T (t)
are aggregated into a model of the power balance deviation in
the open loop mode dPot (t) that is described by the Gaussian
sum [11], [12] with the probability density function:
fGs (dPot ) =

X
i

P (t)
i
e
2dPo (t)

[dPot dP (t)Ti ]2
o
(t)
2 2
dPo

(6)

where Pi (t) are probabilities of the outages and all their

possible combinations and Ti are appropriate amplitudes. The
sum is used for determination of the overall positive and
negative volumes of ancillary services R+ (t), R (t) under
a required reliability level of the system operation. The level
is given as a percentage share of dPct (t) amplitudes larger
than 100M W along all of its possibly existing amplitudes
(the share is called as Value at Risk V aR):
V aR
V aR
), R (t) = FGs (
). (7)
2
2
Since an analytical formula of a cumulative distribution function of the Gaussian sum FGs is not known, R+ (t) and R (t)
have to be computed by a numerical algorithm.
R+ (t) = FGs (1

E. Decomposition of R(t) and economical optimization

The overall volumes are then divided into power reserves
of particular ancillary services on behalf of statistical and
dynamical characteristics of specific components involved in
dPot (t) that should be decayed by corresponding types of
the services (a random component secondary frequency
and power control, a direct - energetic component tertiary
control, a forced outage quick - start, a long - time power
imbalance non - spinning stand - by reserve, emergency
assistance, control energy purchase, etc.).
Finally, the power reserves are economically optimized. The
reserves are re - disposed by the optimization in such a way
that the required reliability level (V aR) is kept and costs to
their purchase are minimized at the same time.
F. Monte-Carlo simulation
In order to make sure that the whole optimization procedure
of the system operation planning is well done, a Monte - Carlo
simulator is employed to verify that the power reserves of
particular ancillary services are large enough for guaranteeing
the required reliability level of the system operation. The
simulator models and simulates electricity transmission system
operation in the closed loop mode. This means that it involves
models that constitute the open loop (model of undisturbed
system operation, model of forced power unit outages) as
well as models that constitute the feedback (models of the
ancillary services mentioned above). The simulator works then
under saturation limits of the services that are given by their
corresponding limited power reserves.
Remark: The approach/method presented in the paper has
been developed for the TSO of the Czech Republic - CEPS a.s.
since 2005 in terms of the Center for Applied Cybernetics. It
is routinely used by the Czech TSO for electricity transmission
system operation planning. Therefore, all results, which will
be presented here, have been obtained through analyses and
simulations of the Czech electricity transmission system operation. Nevertheless, the approach/method is general enough in
order that it is applicable to any other areas. As a consequence,
appropriate results and conclusions obtained can also be considered as sufficiently prevailing despite the approach/method
being developed for the concrete area and its TSO.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
24

III. W IND SPEED AND POWER

A. Wind speed variability
Electrical power produced by wind power stations is dependent on time - position wind speed variability. Position wind
speed variability is given by landscape articulation. Whereas,
time wind speed variability is assessed by changes of weather
conditions. Moreover, wind speed variability is also influenced
by various random fluctuations that represent chaotic wind
behavior. Therefore, it is quite useful to describe wind speed
by its statistical characteristics mean and standard deviation.
Values of these characteristics depend on a position of being
situated. In other words, they depend on longitude, latitude
and altitude values, respectively. Values of the characteristics
for the area of the Czech Republic are illustrated on Fig. 2, 3.

power P (t) and the speed u(t) are supposed to be time varying. This means that they can vary within a specified time
horizon of production. Subsequently, the formula (8) can be
used for determination of mean and standard deviation of wind
power produced by a wind power station whose position is
given by specification of its longitude, latitude and altitude:
Pm =

1
u3 , u3 = (um + ustd )3
2

(9)

where Pm is mean of wind power produced by a wind power

station within a specified time horizon of production, um is
mean of wind speed within the horizon and ustd is its standard
deviation within the horizon. The coefficients , are known.
Pstd =

u
3
ustd u3
, u
3 =
PD
0.35um

(10)

where Pstd is standard deviation of wind power and the

coefficient P D is also known. The coefficients , and P D
and the constant 0.35 have been identified from data.
Consider a wind power station whose location is marked
on Fig. 2, 3. Curves of mean and standard deviation of wind
speed and corresponding mean and standard deviation of wind
power produced by the station are shown in Fig. 4, 5, 6, 7.
11
day horizon
month horizon
year horizon

10.5

Fig. 2.

Mean of wind speed

10
9.5

[mps]

9
8.5
8
7.5
7
6.5
6

1000

2000

Fig. 4.

3000

4000
5000
[hours]

6000

7000

8000

Mean of wind speed

4.4
day horizon
month horizon
year horizon

4.2

Fig. 3.

Standard deviation of wind speed

4
3.8

[mps]

3.6

B. Wind power variability

3.2

As a matter of course, wind power produced by a wind

power station and speed of wind, which blows around the
station, are closely associated with each other. The formula
that connects wind power produced by a wind power station
and wind speed around the station is the following:
1
P (t) = u3 (t)
2

3.4

(8)

where P (t) is wind power produced by a wind power station,

is an efficiency coefficient of a wind power station, is
air density and u(t) is wind speed. The coefficient and the
density are assumed to be constant. On the other hand, the

3
2.8
2.6

Fig. 5.

1000

2000

3000

4000
5000
[hours]

6000

7000

8000

Standard deviation of wind speed

Remark: Values of means and standard deviations of wind

speed and wind power are periodic with a time period that is
equal to 1 day (24 hours) in terms of the day horizon. The
reason is that the values are about consistent for the same
hours through particular days of a month.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
25

1.8
day horizon
month horizon
year horizon

1.6

1.4

[MW]

1.2

0.8

0.6

0.4

0.2

1000

2000

Fig. 6.

3000

4000
5000
[hours]

6000

7000

8000

Mean of wind power

0.45
day horizon
month horizon
year horizon

0.4

0.35

[MW]

0.3

0.25

0.2

0.15

0.1

0.05

Fig. 7.

1000

2000

3000

4000
5000
[hours]

6000

7000

8000

Standard deviation of wind power

a given setpoint up to 30 minutes) and decreased demands on

needs of positive tertiary control. The demands are especially
laid on the negative tertiary control. The reason is that actual
power production of base - load power units has to be restricted
due to random extra power production of wind power stations
so that power balance equilibrium within electricity transmission system operation under 50Hz frequency is guaranteed.
Unfortunately, these restriction procedures result in life cycle
reduction of power units whose actual power production is
restrained in such a way.
V. C ONCLUSION
The paper presents relationship between production capacity
of wind power stations and needs of ancillary services. It
is demonstrated here that extra power production of wind
power stations puts increased demands on power reserves of
certain types of ancillary services. Consequently, this implies
that financial costs of electricity transmission system operation
with wind power stations involved are higher in comparison
with the operation without any wind power station plugged in.
Furthermore, stability of the system with wind power stations
gets worse because their power production level is hardly
ever predictable and this induces that electricity transmission
system operation gets more control - intensive.
ACKNOWLEDGMENT

IV. N EEDS OF ANCILLARY SERVICES

As a matter of fact, extra electrical power produced by
wind power stations and delivered into electricity transmission systems while wind blows influences overall volumes
of ancillary services R+ (t) and R (t) defined by (7). The
volumes appear larger in comparison with that case when the
power is not considered in terms of standard computations.
The reason is that means and standard deviations of the power,
which becomes a part of the power balance deviation in the
open loop mode dPot (t), are included in means and standard
deviations of this quantity whose mathematical model is given
by the Gaussian mixture (6). Consequently, power reserves of
certain kinds of ancillary services also get larger. This fact is
illustrated with the help of values in the following tables:
1) take a wind power station with installed power 1 GW:
SC
T C+
T C
+11% -54% +190%
2) take a wind power station with installed power 2 GW:
SC
T C+
T C
+42% -74% +479%
It is clear from the values mentioned in the tables that
wind power stations with their extra power production cause
increased demands (in comparison with standard electricity
transmission system operation without any wind power station
built) on needs of secondary control - SC (a spinning ancillary
service working in an automatic mode that responds to a given
setpoint up to 10 minutes), increased demands on needs of
negative tertiary control - T C (a spinning or non-spinning
ancillary service working in a dispatcher mode that responds to

The work was supported by the Ministry of Education of

the Czech Republic under the project No. 1M0567.
R EFERENCES
[1] T.W. Anderson, The Statistical Analysis of Time Series, John Wiley &
Sons, New York, 1971.
[2] A. Papoulis and S.U. Pillai, Probability, Random Variables and Stochastic
Processes, McGraw-Hill, New York, 2002.
[3] G.R. Chen and S.H. Hsu, Linear Stochastic Control Systems, CRC Press,
London, 1995.
[4] L. Arnold, Random Dynamical Systems, Springer-Verlag, Berlin, 1998.
[5] E. Janecek, V. Cerny, A. Fialova, and J. Fantik, A New Approach to
Modelling of Electricity Transmission System Operation, Proceedings of
the IEEE PES Power Systems Conference & Exposition, Atlanta, Georgia,
USA, October 29 November 1, 2006, 1429 1434.
[6] P. Havel, P. Horacek, and J. Fantik, Optimal Planning Strategy of Ancillary Services Ensuring Secure Operation of Power System, Proceedings
of the IEEE International Conference Powertech, Lausanne, Switzerland,
July 1 5, 2007.
[7] V. Cerny, P. Janecek, A. Fialova, and J. Fantik, Monte-Carlo Simulation
of Electricity Transmission System Operation, Proceedings of the 17th
IFAC World Congress, Seoul, South Korea, July 6 11, 2008, 4588
4593.
[8] P. Havel, P. Horacek, V. Cerny, and J. Fantik, Optimal Planning of Ancillary Services for Reliable Power Balance Control, IEEE Transactions
on Power Systems, Vol. 23, No. 3, August 2008, 1375 1382.
[9] K. Maslo and S. Vnoucek, Modelling of Steam Unit Outages, Proceedings of the 6th International Conference on Probabilistic Methods Applied
to Power Systems, Madeira Island, Portugal, September 25 28, 2000.
[10] R. Billinton and Y. Li, Incorporating Multi-state Unit Models in Composite System Adequacy Assessment, Proceedings of the 8th International
Conference on Probabilistic Methods Applied to Power Systems, Iowa
State University, Ames, USA, September 12 16, 2004.
[11] H.W. Sorenson and D.L. Alspach, Recursive Bayesian Estimation Using
Gaussian Sums, Automatica, Vol. 7, No. 4, 1971, 465 479.
[12] D.L. Alspach and H.W. Sorenson, Nonlinear Bayesian Estimation Using
Gaussian Sum Approximations,
IEEE Transactions on Automatic
Control, Vol. 17, No. 4, 1972, 439 448.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
26

" ##

$%

*
%
!

&'

%
+

,
%!

'/
( ) '

"#$$$

( )

+
.
0

&

70 "

'$
'$$

2%

' 4

#
% #
0

%
&
%
!

% %

0 *

%
*

#
%
#
4

%
!

- 5
%
-

%
%

2/

#
:;
%

%
%
%

)
6#

'
'

%
&

%
'

'

4
'
%

&
# 0 %
#
0
%

%
:;;
&

2%
#

9
-

0 2%

%
% #

- 02
0

/*

' 4

%
%

--

, "/

%
4
' *

%
#
%

%
#

% 4

)
%

23

% %

#
%
%

% %

0 2%
%

%
-

'

1+ 2

#
#

% -

- #

2%

%
&

%
%

'
%

%
#

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
27

4
%
0

:
<

7=

7>

7?

:;
7
=

7:

:0 *

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
28

A0 1 2 8 ""
79 *

:9
<9 @
>9 1

779
5

7:9 1
7<9 (
7>9 A

B9 4
=9 1
?9 /

&
%

7C9 /
7D9 8 & #
:;9

2%

%
0 *
2% *
E
E'

'
%

%
% % %
#
%

%
%

&
%
% %'

70;
70B
% % % 0 2%

&

%
%
' %

0
%
0

% %

#
-

% 0 2%
0

'
-%
%

7;

0
F

%
%
%
% #
0 2%
'
& % #
' %
-0 2%
0 2%
%
#
%
%
% # # %

'

% % %
%
%
0

-'

0
% %
#

' %
62%
#

'
% %

%
%

1%
%

0 2%
&

2%

0 2%

--

-%

%
#

"" 8(

@
% %

%
%

'

% 4
% %
'
#
#

7B9 @
#
7=9 *
7?9
%

C9 2
D9
7;9 1

%
-

#6 &

% % %
%
% )
' %
% % %
'
% % %
4
#
0 2%
%
#
% 4

%
- 0
%

'

#
# #

%
%
%
0

<0 *
>0 1

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
29

&

4
%
!
#
%

%
%

0 1%

%
' %

%
%
#

%
-'

0 2%

%
%

&'
#

#
0 2% %
#
%
#
% -%
% %%
% %
%
%
% -% )
0 2%
- %
% % %
# - #
%
%
%
0 2%
#
!
# #
%
% % %
#
% 0 2%
%
%
% -#
% %
#
% %
%
&0 2% #
%
% - % %
#
%
%
#
#
) 0
%
%
% #
# % '#
# )
- %
# %
'
#
0
%
% 0
%
%
% - - G
G
#
#
'
#
%
% % %
0

% %
%

%
0 1
% -

( )
% %

0
!

%
%

5#

% - !
B&
% % 4
<:;; 4
% C
'

%
4

%
#
%
-

% #
% %

#
0 :6 %
%

% - !
#
%
% % %
%
%
/
%
4
( ) ' 7;; 4
-%
%
7;; 4
% %
%
%

%
-

'
%
'
%
-

&

A0

% %
#
%

'$ %
%
7 %
% 4
%
&
% %
%
% 4
% %
#
7 #
% %
0
6 % %
-- #
#
#
0
&
>;;
!
% 4
' %
%
7;
' % %
% 4
%
% ' %
#
# 0
% 76 %
7&
'#
%
&
4
#
%
%
%
#
>B
%
4
% - ! '
% % %
%
# % %
% C
% 4
( ) 0
<:;; 4
% %
>B
%
%
#
0
%
#
'
# %
0
.

", .

# &
%
%

%
%
0 2%

- !
E@
#

'
9

--

%
&
-

% 4
%

0 :6
#
% %

% %
%
76
%

%
'
%

%
%

&

&
% %
'
%

#
#

%
4

%
%

% %
4

2 2,8

%
+

% #
'
%
% %

4
-

A0 *

%
' %

%
4
0 2 &
#
#
#
$ '#
% C
0 2%
% 7;
7B;
( ) ' % %
#
%
%
4
( ) 0
%
%

%
%
% -

'

% #

% %
%

# 9 %
% %
%
%

#
- %

&

0
%
% #
%

-!
-

%
%
%
0

%
-

&
1%

%
%

%
%

%
%
%

'

0
' #
& %

&
2%

#
%

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
30

%
%
#
0

:0

70

' %

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
31

Fault Liability and Maintenance Cost Modeling

of a Steam Turbine
Houdova Lenka, Jelinek Libor, Janecek Eduard

Department of Cybernetics, University of West Bohemia

Univerzitni 8, 306 14, Plzen
Pilsen, Czech Republic
Abstract - This paper focuses on modeling of power stations
technological equipment reliability. There is described a method
of failure model creation, based on failure and repairs data of a
particular components, and cost model creation, based on costs
specification. At first there is mentioned the data processing
procedure, problems of appropriate stochastic parameters
inference and total characterization of failure model connected
with small number of data and their dependence on operational
conditions in concrete. This model should be able to predict the
future conditions and to estimate important evaluation indicators.
For that reason it is necessary to introduce the software, which is
used for simulation, visualization and optimization future
development of evaluation indicators possibilities. This work is a
part of the joint project of the UWB and koda Power company
called Operational availability and maintenance cost
optimization with the computer support.

I.

INTRODUCTION

One of the basic tasks of management in operational

reliability control area is planning and making the strategic
decisions. It is important especially in energetic, where
an effect of maintenance prove at fault liability and
expensiveness of potential consequences. Planned and
preventive maintenance activities are instruments for
technological equipment lifetime raising and for contractual
power supply provision. These activities have decisive effect to
total operational cost too. In order to be optimal this service
task, it is necessary to make an extended analysis, failure
causes and effects calculations with respect to task effect and
production progress strategy.
The aim of this paper is to bring out the problem
of processing the information from steam turbine operation to
proper form for effective maintenance planning especially
based on cost evaluation of outage risk and maintenance.
The principle is the creation of the probability model
of operational condition from historical data. This model
should be able to predict the future conditions and to estimate
important evaluation indicators. With the aid of input
conditions and simulation parameters changes it is possible to
make sensitivity analysis of indicators in reliability interval,
iterative optimizations and finding of the best strategic solution
according to given criteria e.g.
II.

liability character of steam turbine realization in accordance to

component failure and repairs data, obtained from koda
Power (producer). The main aim is time development
monitoring of technological unit fault liability. According the
analysis there was find out, that the fundamental problem is
finding so-called hidden failure model of components without
planned maintenance execution (left scheme) based on failure
data of components at real operation without applying the
planned maintenance (right scheme), see fig. 1.

DATA PROCESSING

A. Data processing basic themes

Model design stem from demand after the components fault

Figure 1. Scheme of failure model with and without planned maintenance

Reduction of planned maintenance effect is possible to solve

with expert based adjustment of failure data set or model
parameters. To components, which was called for repair at
planned maintenance, is attached a failure. Further way is to
take a presumption, that failure rate will be rise after the time
of planned maintenance interval. The modified model has more
pessimistic character then the original, but it is more realistic.
Other problem is definitely small number of failure data
of selected component with the low or zero observed fault
liability and high technical lifetime. In the case, it is not
possible to approximate more complicated stochastic functions
by ordinary way. Therefore complex data processing must be
used with respect to knowledge based presumptions.
These presumptions are:
a) time progression of components failure rate matches to
bathtube curve,
b) mean failure rate can be determined from designed
technical lifetime.
According to the assumptions is possible to set a shape of the
estimation parametric function curve or to extend a statistical
failure data.
B. Statistical analysis
For stochastic probability distribution of failure origination
special parameters must be defined. These can be taken from
producer of components or from some reliable database (only

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
32

if there was guaranteed an identical form of definition,

comparable construction and components properties). However,
mostly this information is not available. Then it is necessary to
use a standard statistical solution ([1],[2],[3]). According to the
historical data about time duration between failures and repair
duration it is possible to estimate needful parameters by data
analysis. For statistical parameters estimation it is necessary to
have a sufficient number of correct values obtained by many
realizations. It follows, that the sparse data is necessary to put
together for resultant estimation credibility. For compounding
of components is used a bunching algorithm repeatedly,
orientated firstly to similarity of particular components failure
frequency histogram and then to mean time of repair.
McQueen bunching algorithm (so-called k-means) is used for
implementation. After data compounding it is already possible
to estimate standardized group parameters of failure
characteristics, which can be later linear recalculate for
particular component parameters (with using predefine
particular mean time between failure).
For fault liability of mechanical parts definition it is most
often applied Tri-weibull stochastic function [2]:

( x) =

1 x 1 1 2 x 2 1 3 x 3 1
.
+
+
11
22
33

(1)

It takes into account time progression of failure rate because of

aging very well (otherwise bathtube curve). Module Weibull
analysis of Availability Workbench (AWB) software by
company Isograph is used to bathtube curve forming. Needful
parameters a characterizing the interlay curve are
obtained by cumulative probability graph after particular data
groups import.

On the figures 2 and 3 is displayed a standardized

progression of given group failure rate. The results of
automatic parameters estimation for particular groups need not
to be always satisfactory, that is why it is necessary to adjust
them by hand for proper form of failure rate curve (according
to expert estimations). After standardized parameters
recalculation into particular components parameters is obtained
sufficient information for components failure rate progression
modeling. For recalculation there is used the time
transformation. The basic criteria for this transformation is the
curve shape conservation and its proportional expansion to real
values of mean time between failures computed from data. The
time transformation is applied to Tri-weibull distribution
parameters of each component:

='
= ' MTTF

(2)

where and are original parameters of standardized

distribution for component groups and MTTF is mean time to
failure of particular component.
III.

AVAILABILITY AND COST MODEL

A.. Model creation

One of the basic efficiency indicators of steam turbine
energy operation is availability, from which is possible to
calculate specific technical fault liability. Time progress can be
obtained by simulation according to real operation behavior.
For implementation of availability model there is used module
AvSim of software Availability Workbench.
This module offers the tools for definition and editing
of stochastic models string data structure in a several most
widespread formalisms and hierarchical structured levels
of components characterization to the component cause of
failure level.

Figure 2. Example of Cumulative probability graph

Figure 4. Hierarchical structured levels of components

Figure 3. Example of Failure rate graph

In model structure point of view, there is used serial

sequence principle of selected critical components reliability
blocks (the particular components operational probability
multiply themselves). Decisive aspect is certainly failure cause
and effect ([4]).
To model there can be assigned a various additional
information and operational conditions including of costs
specification, which are fundamental for cost model. This
model is created in other module of AWB called RCMCost.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
33

Simulation is proceeded by Monte-Carlo ([5]) method. There

are discerned between the operational and failure condition.
Availability is determine as relative ratio of operation duration
and total time in view. A number of realizations of random
processes are generated and resultant condition is computed as
the mean.
B. Planned maintenance effects on fault liability modeling
For every component there exists parameters set, which
represents failure origin probability distribution in Tri-weibull
function form. Appropriate failure rate function representation
is the best way how to characterize the risk progress of failure
origin in time on similar conditions and presumptions, under
them was the function derived from real failure data.
Due to the information of the major part of the components
technical lifetime the parameter estimation is achieved, which
eliminates planned maintenance effect from data (at least
in part). This effect can be modeled independently (in the
various times).
Effect of planned maintenance to the failure rate
development is so-called age reduction from probability aspect,
which developed a reduction of effective lifetime of every
component and whole modeled equipment.
The failure rate progress is changed by the planned
maintenance implementation to the model in accordance to the
original data. The simulation results should agree with the real
data.
Figure 5 shows the example of common component failure
rate progress in calendar time without planned maintenance,
when failure rate curve follows up the bathtube curve and with
planned maintenance, which is made every 5 years.

of fault liability, which can be analyzed from data. Mean

of years availability relating to whole monitored operational
term is chosen as the verification characteristic. By the ratio
of real availability (calculated from data) and simulation
availability (estimated by model on the same conditions) the
fault coefficient is determined. This fault coefficient is used for
proportional optimization of the correction coefficient
in accordance to statistical relevancy of number of data.
D. Model correction
The aim of correction is the linear reduction of the failure
rate function value. Coefficient of correction depends on
number of data for each component and it can be computed
in accordance to formula:
K=

Q+ N
1+ N

(3)

where Q is optimization parameter, which is changed by Q

in every iteration, and N is real failure number of data, from
which is estimated given parameter set. This computation is
based on presumption, that unavailability increase because
of added components failures is necessary for unavailability
reduction compensation, especially for components from small
number of data.
Model parameters correction is based on transformation
of stochastic distribution coefficients. In the case of
exponential distribution the simplest way is multiplying by one
parameter K mean time to failure with correction:
* ( x) =

( x)
1
,
=
K
K MTTF

(4)

In the case of Tri-weibull distribution the situation is more

complicated. It is necessary to multiply every term
denominator, then total failure rate become reduced
independently from time and curve form is saved. It is
appropriate to make correction just for coefficients in
denominator. Calculation is made in accordance to formula:
* ( x) =

1 x 1 1

(K

11

11 ) 1

2 x 2 1
(K

21

2 2 ) 2

3 x 3 1
1

( K 3 33 ) 3

(5)

Result of iterative optimization is parameters of particular

components failure model set correction.
IV.
Figure 5. Example of common component failure rate

C. Model verification
Because of small number of data reliable the stochastic
distribution parameters estimations arent fully. Thats why it
is not possible to consider the result of first simulations to be
the final result. It is important to set up the common model so
that the simulations results could be comparable with real
condition of concrete equipment.
Due to presumptions and hypothesis it is necessary to
specify the model according to some stable characteristic

AVAILABILITY AND COST SIMULATION, OPTIMIZATION

For getting the close results after model correction and

verification it is necessary to correct and completely
characterize the availability and cost model from all
perspectives (labour, spares, equipment, failure effects, prices
etc.). Used software AWB has strictly defined needed
resources and functionalities. The scheme of resources and
functionalities can be seen on the figure 6.
Simulations made by AWB are focused on fault liability
indicator development location. There are randomly generated
failure events in defined time horizon and number
of realizations according to the required parameters.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
34

Figure 8. Principal of cost optimization

It is very important to execute the sensitivity analysis, which

can show the potential range of resultant characteristics for
selected availability parameters estimation interval. Sensitivity
analysis can be taken for relative change of mean time to
failure of particular components, mean time of repair or for age
reduction factor (it characterizes the quality of repair) changes
see fig.9.

Figure 6. Scheme of needful resources and functionalities of AWB

The simulation results are mean values of availability in

every particular period (1 year e.g.) in simulation horizon.
According to project assignment the attention is concentrated
on simulation of the history and prediction of availability and
specific technical fault liability. Many parameters are set up or
adjusted according to the expert estimations. On the figure 7
there is shown running of the specific technical fault liability in
history.

Figure 7. Common steam turbine availability simulation (history)

Cost simulations and planned maintenance interval

optimizations are made in RCMCost. Parameters of failure
origin probability distribution for cost model are imported from
AvSim module and additionally there are rated by cost every
maintenance units and failure effects. Software enables the
optimal interval calculation from the cost and from the
availability perspective.
Simulations are aimed at general and ordinary maintenance
interval cost optimization. Optimal interval is a compromise
between maintenance costs and costs to repair after failure (see
fig. 8).

Figure 9. Sensitivity analysis of age reduction factor (ARF)

V.

CONCLUSION

Fault liability modeling and simulations of condensing steam

turbine described above is focused on availability and cost
prediction (with the option to change the operational condition,
particular terms and planned maintenance range). As the results
it can be computed for example expected operational
availability values, which could be used for power supply
provision guarantee or as recommendation for planned
maintenance optimal time and range. These resultant technicaleconomic indicators are useful for turbine producers, which
must guarantee the failure-free operation and service, and for
users (energy companies), which can predict their own
management conditions and the expansion planning.
REFERENCES
[1]
[2]
[3]
[4]
[5]

J. Novotny, The Steam Turbine 200 MW Reliability, Diploma work: ZCU

Plzen, 2008, in Czech only
A. Papoulis, Probability, Random Variables, and Stochastic Processes,
4th ed., Boston, 2002
V. Kuchar, Methodology of Reliability Data in UE CEZ a.s. Evidence,
I&C Energo Brno: Brno, 2006, in Czech only
Isograph, AWB 2007 user guide, Isograph Limited, 2007
M.A. Tanner, Tools for Statistical Inference, 3rd ed., Springer Verlag:
New York, 1996

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
35

VOLTAGE STABILITY ANALYSIS OF GRIDS CONNECTED WIND

GENERATORS
1

Nguyen Tung Linh 2 ;Trinh Trong Chuong;

Electric Power University

Ha Noi University of Industry

Abstract
Wind generators can have a significant impact on the power flow, voltage profile and the power quality for
customers and electricity suppliers. This requires a suitable tool to analyze the influence of wind generators on the
distribution system. This paper presents a method to find the steady-state voltage stability region for each bus of a
distribution power system, considering the presence of windpower generation. The maximum permissible load of
each bus is calculated, so that it can operate with the voltage within the limits allowed by the power system utilities.
Keywords: Wind generator, Voltage stability

1. Introduction.
Recently wind power generation has been experiencing
a rapid development in a global scale. The size of wind
turbines and wind farms are increasing quickly; a large
amount of wind power is integrated into the power
system. As the wind power penetration into the grid
increases quickly, the influence of wind turbines on the
power quality and voltage stability is becoming more
and more important. It is well known that a huge
penetration of wind energy in a power system may cause
important problems due to the random nature of the
wind and the characteristics of the wind generators. In
large wind farms connected to the transmission network
(110 kV 220 kV) the main technical constraint to take
into account is the power system transient stability that
could be lost when, for example, a voltage dip causes
the switch off of a large number of wind generators.
In the case of smaller installations connected to weak
electric grids such as medium voltage distribution
networks, power quality problems may became a serious
concern because of the proximity of the generators to
the loads. The existence of voltage dips is one of the
main disturbances related to power quality in
distribution networks. In developed countries, it is
known that from 75% up to 95% of the industrial sector
claims to the electric distribution companies are related
to problems originated by this disturbance type. These
problems arise from the fact that many electrical loads
are not designed to maintain their normal use behaviour
during a voltage dip. The aim of this paper is to conduct
a voltage stability analysis using an iterative power
system simulation package, to evaluate the impact of
strategically placed wind generators on distribution
systems with respect to the critical voltage variations
and collapse margins. This paper concludes with the
discussion of wind generators excellent options for
voltage stability.

Induction machines are use extensively in the power

system as induction motors but are not widely used as
generators. Despite their simplicity in construction, they
are not preferred as much as synchronous generators.
This is mainly due to the defined relationship between
the export of P and absorption of Q. However, induction
generators have the benefits of providing large damping
torque in the prime mover, which makes it suitable for
the application in fixed speed wind turbines. The fixed
speed wind turbine uses a squirrel cage induction
generator that is coupled to the power system through a
connecting transformer as shown in Figure 1. Due to
different operating speeds of the wind turbine rotor and
generator, a gearbox is used to match these speeds. The
generator slip slightly varies with the amount of
generated power and is therefore not entirely constant.
Wind turbine
gearbox

G
Induction
Generator

vG PG QG
iG

Grid

Figure 1: Modelling wind turbine connected grid

However, because these speed variations are in the order
of 1 per cent this wind turbine is normally referred to as
constant speed. Nowadays, this type of wind turbine is
nearly always combined with stall control of the
aerodynamic power, although pitch-controlled constant
speed wind turbine types have been built in the past.
Induction machines consume reactive power and
consequently, it is present practice to provide power
factor correction capacitors at each wind turbine. These
are typically rated at around 30 per cent of the wind
farmcapacity. As the stator voltage of most wind turbine
electrical generators is 690V, the connecting transformer
of the wind turbine is essential for connection to the
distribution network and should be considered when

2. Induction machines.
8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
36

modeling the electrical interaction with the power

system [3].
3. Impacts of WGs.
Connecting a generation scheme to a distribution
network will affect the operation and performance of
the network depending on the scheme and rating of the
generator itself. The impacts are as follows:

Power Flows
Voltage stability
Fault Analysis
Impact of WTs on the Networks

voltage drops rapidly when there is an increase in the

load demand. Load flow solutions do not converge
beyond this point, which indicates that the system has
become unstable. This point is called the Critical point.
Hence, the curve can be used to determine the systems
critical operating voltage and collapse margin.
Generally, operating points above the critical point
signifies a stable system. If the operating points are
below the critical point, the system is diagnosed to be in
an unstable condition [5].

3.1 Voltage Stability

A system experiences a state of voltage instability when
there is a progressive or uncontrollable drop in voltage
magnitude after a disturbance, increase in load demand
or change in operating condition. The main factor,
which causes these unacceptable voltage profiles, is the
inability of the distribution system to meet the demand
for reactive power. Under normal operating conditions,
the bus voltage magnitude (V) increases as Q injected at
the same bus is increased. However, when V of any one
of the systems buses decreases with the increase in Q
for that same bus, the system is said to be unstable.
Although the voltage instability is a localised problem,
its impact on the system can be wide spread as it
depends on the relationship between transmitted P,
injected Q and receiving end V. These relationships play
an important role in the stability analysis and can be
displayed graphically [1].
PV Curves
When considering voltage stability, the relationship
between transmitted P and receiving end V is of interest.
The voltage stability analysis process involves the
transfer of P from one region of a system to another, and
monitoring the effects to the system voltages. This type
of analysis is commonly referred to as a PV study.

Figure 3: Charatic Q-V

QV Curves
Voltage stability depends on how the variations in Q and
P affect the voltages at the load buses. The influence of
reactive power characteristics of devices at the receiving
end is more apparent in a QV [Link] shows the
sensitivity and variation of bus voltages with respect to
reactive power injections or absorptions. Figure 3 shows
a typical QV curve, which is usually generated by a
series of load-flow solutions. Figure 3 shows a voltage
stability limit at the point where the derivative dQ/dV is
zero. This point also defines the minimum reactive
power requirement for a stable operation. An increase in
Q will result an increase in voltage during normal
operating conditions. Hence, if the operating point is on
the right side of the curve, the system is said to be
stable. Conversely, operating points in the left side of
the graph are deemed to be unstable [4, 5].
PQ curves.

The maximum permissible loading of a system, within

the voltage stability limit, is usually determined from the
well known PV curve or QV curve. The PV curve is
plotted for a constant power factor and the QVcurve is
plotted for a constant power. A series of computer
simulations is required to generate a family of these
curves [6]. There are other methods to find the voltage
Figure 2: Charactic P-V
stability limit of a system, such as multiple load flow
solutions, singularity criterion of the Jacobian matrix
The Figure 2 shows a typical PV curve. It represents the [7]. It is an established fact that the voltage collapse
variation in voltage at a particular bus as a function of occurs when the system load increases beyond a certain
the total active power supplied to loads or sinking areas. limit. If the limiting values of P and Q are known, the
It can8. EEEIC
be seen
that at the knee of the PV curve, the voltage
stability
margin
for a May
given
operating
International Conference on Environment and Electrical
Engineering,
Karpacz,
Poland,
10-13,
2009 point can
37

directly be determined. This requires the plotting of

voltage stability boundary of the system in PQ plane,
however, using the limiting values of P and Q. To the
best knowledge of the author, no such work has been
reported so far that can determine the voltage stability
margin using the PQ curve.
In this letter, the limiting or critical values of P and Q at
the voltage collapse point are first determined and then
used to plot the voltage stability boundary in PQ plane.
Unlike the conventional PV or QV curves, no fixed
value of power or power factor is used in generating the
stability boundary [8]. Using the above curve, the
voltage stability margin in terms of P, Q, or S (for a
given power factor) can easily be determined when the
initial operating point is known. Consider the equivalent
model of a distribution line connected between bus i
and bus j as shown in Fig.4.

a = c1 = a12 + a22 ; b = c2 Pijr + c3Qijr Vi 2 ; c = c4 ( Pijr2 + Qijr2 )

c1 = a12 + a22 ; c2 = 2(a1b1 + a2b2 ); c3 = 2(a1b2 + a2b1 )
c4 = b12 + b22
The solution of (5) is the square of the receiving end
voltage. Thus the receiving end voltage can be
calculated from (6).
Vj =

b b 2 4ac
2a

(6)

There are two solutions for (6), the lower solution lies
on the lower part of the P-V curve and is unstable [4].
Thus, the available solution is a stable one on the upper
half, which can be expressed as in (7).
Vj =

b b 2 4ac
2a

(7)

The point where the two trajectories, i.e. stable and

unstable lines, are joined is the nose or bifurcation point.
In a dition, this is the point where the maximum power
can be transferred, which is the condition b 2 4ac = 0 .
Substitute coeffcients of the quadratic equation from (4)
into (7) and rearrange, we obtain (8).
( c 22 4 c1 c 4 ) Pijr2 + ( c 32 4 c1 c 4 ) Q ijr2 2 c 2V i 2 Pijr

Fig.4: The model distribution line connected between

bus i and bus j
We can formulate the relationship between injected
current and voltage at any buses based on generalized
ABCD parameters as follows:

Vi i = AVi i + BI j

2 c 3V i 2 Q ijr + 2 c 2 c 3 Pijr Q ijr + V i 4

(8)
The relationship between Pijr and Qijr of (8) is a locus of
the collapsing point on the P-Q plane which separates
the operating points into feasible and infeasible regions.
The calculation method to obtain the predicted
collapsing point can be illustrated in Fig.5.

(1)

where A = 1 + ZYc and B = Z. The complex form of A

and B can be expressed as shown in (2).
A = a1 + ja2 and B = b1 + jb2

(2)

The receiving end current, Iijr, can be expressed:

I ijr = (Pijr jQijr ) / Vi i

(3)

Substitute A and B from (2) and Iijr from (3) into (1)
resulting in (4).
(b1 + jb2 )( Pijr jQijr )
Vi i = (a1 + ja 2 )V j j +
V j j
(4)
or:

c1V j4 + (c 2 Pijr + c 3 Qijr Vi 2 )V j2 + c 4 ( Pijr2 + Qijr2 ) = 0

4

or: aV j + bV j + c = 0
where:

(5)

Fig.5:The predicted collapsing point calculation.

3.2 Test Systems.
Fig. 6 shows the simplified diagram of test system. It
has 20 buses operating at the voltage of 12.47 kV, 19
circuits/transformers, and the hypothetical wind power
is connected at bus 20 at 0.69 kV. The detailed of the
system is not showed but it has only one circuit of 138
kV [4]. The power wind farm is a variable speed
directdrive synchronous generator connected to the
Point of Common Coupling, (PCC) by a full-load
converter. It can be represented by a P-Q bus or a P-V
bus, since it has the capacity to control the reactive
power.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
38

Table 1. LINE PARAMETERS OF DISTRIBUTION

SYSTEM IN PER UNIT ON 40 MVA BASE.
Line number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

From
1
2
2
2
3
3
5
5
5
10
10
11
11
12
12
16
16
19
20

To
2
3
9
10
4
5
6
7
8
11
16
12
15
13
14
17
18
1
12

Resistance
0.1196
0.125
0.0713
0.0664
0.0936
0.6294
0.3123
0.624
0.6559
0.0561
0.0859
0.0471
0.037
0.08
0.0257
0.0229
0.0286
0
0

Reactance
0.1263
0.1505
0.0185
0.0702
0.1129
0.7586
0.3766
0.4947
0.5201
0.0592
0.0908
0.0496
0.0095
0.0844
0.0067
0.0059
0.0075
0.28
0.28

First: the wind power generation is connected and

operates as a P-Q bus with unit power factor. His
particular operating point for actual operation state is PG
= 7.5 MW and QG = 0.0 MVAR . The Wind Power can
also be represented by a P-V bus with the reactive power
limits, so that when the limits are reached, it becomes a
P-Q bus. In modern wind turbine system it is possible to
control the power factor, provides in this way, an
additional source of reactive to help the power systems
to control de voltage.
Second: without Wind Power Generation.
Fig. 7, 8 show the first scenario and the second for the
load buses 13 of the system.

TABLE II: DISTRIBUTION SYSTEM BUS DATA

AFTER THE CONVERGENCE OF THE POWER
FLOW.
Bus
number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20

PL,
MW
0.4
0.4
0.4
0.5
0.5
0.6
0.5
0.5
0.8
0.75
1
2
1
0.75
0.75
1.5
0.5
0.5
0
0

QL,
MVAR
0.15
0.15
0.15
0.2
0.2
0.25
0.2
0.2
0.3
0.3
0.4
0.85
0.35
0.3
0.3
0.6
0.2
0.2
0
0

PG,
MW
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
6.3
7.5

QG,
MVAR
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
6.8
0

The loads are distributed considering the thermal limits

of the conductors. The line parameters, loads and
Generator data are at the Tables I and II. The load is
increased with steps of 0.5% in all loads, starting from a
normal operating point. The limits of the safe voltage in
distribution buses are 0.95 to 1.05 per unit. The system
was simulated considering two different scenarios.

Fig.6: Distribution Power System Test.

U, kV
12
10
8.0
6.0
4.0
2.0

P, MW
0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

Fig. 7. Loading in bus 13 - Voltage stability region with

the contribution of wind power.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
39

U, kV

Integrating Large Wind Farms into Weak Power Grids

with Long Transmission Lines; IEEE/PES Transmission
and Distribution Conference & Exhibition, China, 2005.
[7]. L. S. Barros, W. S. Mota, D. F. P. Moura; Matrix
Method to Linearization and State Space Representation
of Power Systems Containing Doubly Fed Induction
Machines Operating as Wind Generators; 2006 IEEE
PES Transmission and Distribution Conference and
Exposition Latin America, Venezuela.
[8] Poul Srensen, Anca Hansen, Lorand Janosi, John
Bech, Birgitte Bak-Jensen; Simulation of Interaction
between Wind Farm and Power System; Ris National
Laboratory, Roskilde December 2001.

12
10
8.0
6.0
4.0
2.0
0

P, MW
0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Fig. 8. Loading in bus 13 - Voltage stability region

without wind power.
4. CONCLUSION
A very efficient method is presented to determine the
maximum permissible load that can be allowed in each
bus of a distribution power system. To ensure the
reliability of the results, a comparison between the
proposed and a previous analytical method is carried out
in a simple two bus system connected by a single
transmission line. The results are very close and the
proposed method was applied on a distribution power
system. The presence of wind power is considered in
order to analyze the effect of this generation on voltage
operation and at the voltage stability limits. The voltage
stability limit region and a sub-region in which the
voltages at the bus are in acceptable ranges are defined
for different power factors. For practical purposes, the
PQ curves can be useful to determine the capacity of
certain buses of the system to support an increasing in
the load demand. Of course, to meet de new demand, the
limits of secure voltages must be determined. P-Q
curves can help the utilities in the choice the best point
to connect the wind farms.
5. References.
[1]. [Link]: Power: System Stability and Control,
New York: McGraw-Hill, 1994.
[2]. Ake Anderson; Power quality of wind turbines;
Thesis Degree of Doctor of Philosophi; Chalmers
Unversity of Technology; Goteborg Swedent 2000.
[3]. Fritz W. Mohn, C. Zambroni de Souza: Tracing PV
and QV Curves with the Help of a CRIC Continuation
Method, IEEE Trans. on Power Systems, Vol. 21, N 3,
pp.1115- 1122, 2006.
[4]. McGraw-Edison Company, Power System Division:
Distribution-System Protection Manual.
[5]. M. Chinchilla, S. Arnalte, J.C. Burgos, and J.L.
Rodrguez: Power limits of grid-connected modern wind
energy systems, Renewable Energy, September 2005,
pp. 1455-1470.
[6]. Richard Piwko, Nicholas Miller, Juan Sanchez
Gasca,
Xiaoming
Yuan,Conference
RenchangonDai,
James Lyons;
8. EEEIC
International
Environment
and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
40

Organic solar cells based on ZnO nanowires layer

Ji Podzemsk1, Wolfgang Schade2
1

Czech Technical University, Faculty of Electrical Engineering

Technick 2
Prague, Czech Republic
2
TU Clausthal, Institute of Physics and Physical Technologies
Leibnitzstr. 4
Clausthal-Zellerfeld, Germany

Abstract- In our research we are looking for alternative to

solar cells based on silicon. We are using zinc oxide (ZnO) which
is N type semiconductor instead of silicon. To reach PN junction P
type polymer is used. ZnO is in form of well aligned nanowires.
Since the distance between nanowires is smaller than the mean
free path of electrons almost all emitted electrons are collected.
We use vapor-liquid-solid (VLS) method to create ZnO nanowires.
Used substrate was sapphire, silicon and silica glass. Small solar
cell was manufactured and its open circuit voltage under light was
measured.

I.

INTRODUCTION

Scientists all around the world are looking for new energy
resources or are improving current techniques of conversion to
electric energy. Conversion from sunlight is one of promising
method. Except sufficient natural conditions, the growing price
of silicon is one of its disadvantages. Conventional solar cell
uses tree or comb structure of contacts to collect emitted
electrons. These structure must be fine enough comparable
with mean free path of electron. If not so, emitted electrons
recombine inside the semiconductor material and are not
generating desired current. One way how to collect all emitted
electrons is to make the structure of contact very fine. In our
case one of contacts is semiconductor zinc oxide in form of
well aligned nanowires. Herewith collecting electrode is inside
the structure of solar cell and makes active part of PN junction.
In addition the structure of ZnO nanowires is very fine.
Zinc oxide is transparent N type semiconductor material with
wide direct band gap 3.4 eV at room temperature. This direct
band gap is profitable for use in solar cell applications. Other
properties of ZnO favorable for electronic applications include
its stability to high-energy radiation and to wet chemical
etching. Radiation resistance makes ZnO a suitable candidate
for space applications (it corresponds to fact that solar panels
are used as a power source) [4].
II. EXPERIMENTS AND RESULTS
There are many methods how to create well aligned structure
of zinc oxide nanowires. We used vapor-liquid-solid method
(see fig. 1) [1]. This method is very sensitive to the substrate
because the principle of this method is based on epitaxial
growth. It is necessary to have substrate with internal structure
very similar to structure of zinc oxide, low lattice mishmash is
required. Such we used silica glass, sapphire and silicon. The

experiments were done also with ordinary lab glass, but the
result was like expected (in means no structure of zinc oxide
was created on it). The lattice mishmash also influences a
structure of created nanowires layer. Sapphire as a material
with internal structure very similar to one of zinc oxide allows
the growth of perpendicular zinc oxide wires. On the other
hand on substrate made of silicon and silica glass the wires
have disordered structure.
In process of manufacturing of zinc oxide nanowires we put
onto glass boat our substrate covered with gold film and near to
the substrate we put mixture of ZnO powder and graphite
powder (see fig. 4). This boat is given into the furnace with
defined heating cycle (maximum temperature is 1100
1200 C). Argon atmosphere was used. ZnO powder is
reduced by graphite to Zn vapor and CO/CO2. The Zn vapor is
transported to substrate and creates alloy with gold. When the
alloy (droplet) is saturated, Zn oxidizes with oxygen and
nucleates under the Zn-Au alloy droplet. Such ZnO nanowire
(with Zn-Au alloy tip) is formed. The position where nanowire
will appear is selected by the presence of gold clusters. Such
we can create pattern of zinc oxide nanowires. The thickness of
gold which we use in our experiments is 2.5 nm. All substrates
had dimensions 10x10 mm.

Figure 1. Basic scheme of the VLS process in a tube furnace [2].

When the maximum temperature was kept for the duration of

desired time the furnace is turned off and the cooling is
provided under the natural cooling. When the temperature
inside is near to the room temperature, the boat with our
substrate is taken out of the furnace.
The diameter of nanowire grown on silicon and silica glass
was similar and was about 100 nm, on sapphire the nanowire
was thicker with diameter about 350 nm (see fig. 2). The
diameter of nanowire on the substrate depends on the thickness
of gold cluster. Thicker gold cluster creates thicker nanowire.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
41

The length of nanowires depends on the time during which

the maximum temperature is applied. The type of substrate
does not influence the length a lot. The time changed from 35
to 60 minutes and the related length varied from 1 to 15 m.
Also the length of nanowires was not constant on the same
substrate. For example on the same substrate a length of
nanowires varied from 7 to 13 m. That is obvious from the
position of the substrate on the boat (see fig. 4).

was dilute with distilled water to reach concentration 1:3, 1:5,

1:8 (polymer : distilled water). Concentrated polymer was
applied in different thicknesses. Different thickness was
prepared such that we put drop of concentrated polymer on the
substrate and we regulated the volume of concentrated polymer
with syringe. Such we sucked out concentrated polymer but the
diameter of created drop was same. That was very easy way
how to control the thickness. When dried the area was same
but the thickness of polymer differed.

Figure 2. SEM image of structure of ZnO nanowires on silicon substrate.

Figure 4. Boat with mixture of zinc oxide and graphite on the sides,
substrate is put in the centre of the boat. White part is reacted ZnO. Situation
after heating cycle substrate is covered with ZnO.

Figure 3. SEM image of structure of ZnO nanowires on sapphire substrate.

The method of VLC growth is very sensitive. The furnace

has to have temperature zones to reach temperature gradient to
allow vapor of zinc to condensate. The result if layer of ZnO is
created or not very depends on the position of our substrate and
position of zinc oxide source inside the tube furnace.
Droplet of polymer (P type semiconductor) is applied on the
surface of substrate covered with ZnO nanowires and when the
polymer is dried polymer and ZnO are contacted with silver
varnish. Such created solar cell is on fig. 5.
Used polymer as a P semiconductor was PEDOT
(commercial name Clevios P by company H.C. Starck) with
chemical
name
poly(3,4-ethylenedioxythiophene)
poly(styrenesulfonate)
aqueous
dispersion.
Different
concentrations and thickness of polymer was tested. Polymer

Figure 5. Realization of solar cell.

Open circuit voltage of created solar cell was measured

under light intensity 45 000 lx. Spectrum and intensity of light
is on the fig. 6 (for comparison sunlight is depictured also).
Reached open circuit voltage for each combination of thickness
of polymer and concentration of PEDOT polymer is shown in
Table I. After measuring open circuit voltage, thickness of
polymer and zinc oxide layer was measured with using atomic
force microscope. Measured values are in Table II.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
42

Table III. Highest open circuit voltage was 0.9 mV and this
was reached for thickness of polymer 2.3 m (thickness of zinc
oxide nanowires was 1.22 m). Such we got highest open
circuit voltage than in case of silicon substrate. We
demonstrated that values gained in case of silicon substrate are
not influenced by the material of substrate because we got
maximum voltage comparable to one on silicon which is N
type semiconductor. Silica glass is insulator.
TABLE III
THICKNESS OF POLYMER AND OPEN CIRCUIT VOLTAGE OF SOLAR CELL
CREATED ON SILICA GLASS. LIGHT INTENSITY 45 000 LX. THICKNESS OF ZINC
OXIDE 1,22 MICRONS.

polymer (m)
OCV (mV)
Figure 6. Spectrum and light intensity of halogen lamp and sunlight in the
laboratory.

TABLE I
OPEN CIRCUIT VOLTAGE OF SOLAR CELL WITH DIFFERENT THICKNESS OF
POLYMER ON SILICON. LIGHT INTENSITY IS 45 000 LX. ALL VALUES ARE IN
MILIVOLTS.

conc.
1:3
1:5
1:8

thickness of polymer
high
medium
low
0.8
0.6
0.1
0.6
0.2
0.2
0.2

TABLE II
THICKNESS OF POLYMER AND ZINC OXIDE LAYER ON SILICON

conc.
1:3

ZnO
poly

1:5

ZnO
poly

1:8

ZnO
poly

thickness of polymer
high
medium
low
7.05
9.4
12.75
2.46
1.51
1.36
high
medium
low
9.87
13.61
9.58
3.5
3.18
1.51
high
13.95
5.08

As shown in the table I, the highest open circuit voltage on

silicon substrate was reached for polymer PEDOT
concentration 1:3 (0.8 mV). That is why we decided to apply
this concentration in case of using silica glass as a substrate. In
case of silica glass we applied 4 types of thickness of polymer
PEDOT in concentration 1:3. Thickness of zinc oxide
nanowires layer was constant in all surface of our substrate and
was 1.22 m. Results of measured open circuit voltage are in

thickness of polymer
high medium 1 medium 2
3.44
2.3
1.83
0.8
0.9
0.6

low
0.99
0.3

PEDOT polymer was also applied on nanowires on sapphire.

There was not measured any voltage. The reason is very simple.
The nanowires are perpendicular and there are not in electrical
contact.
III. CONCLUSION
Organic solar cell based on layer of zinc oxide nanowires
was created. N part of junction was zinc oxide, P part polymer
PEDOT. Maximum reached open circuit voltage was 0.9 mV.
Base material of substrate does not influence gained voltage,
serves only as a substrate for creation of zinc oxide nanowires.
It is not supposed to manufacture solar panels with method
mentioned above (vapor-liquid-solid method) because this
method is limited by material of substrate. Materials which
serve as a substrate for creation of zinc oxide nanowires are
expensive and the process is energy expensive. But with
chemical methods we can create zinc oxide nanowires structure
with very cheap way (cheap substrate, cheap chemicals, low
temperature) [5, 6]. This method could be very promising for
the future. Although the voltage for one cell is not very high,
we are able to manufacture huge low cost solar panels which
could be embodied to roofing or facade rendering of our houses.
ACKNOWLEDGMENT
The authors gratefully thank Farzaneh Fattahi for technical
supports.
REFERENCES
[1]
[2]
[3]
[4]

P. Yang, Controlled Growth of ZnO Nanowires and Their Optical

Properties, in Advanced Functional Materials, vol. 12, pp. 323-331, May
2002.
[Link]
B. Shaijing, C. Chunxiang, W. Quingzhou, and B. Ling, Growth of ZnO
Nanowires in Aqueous Solution by a Dissolution-Growth Mechnism, in
Journal of Nanomaterials, vol. 2008, Article ID 610541.
D. C. Look, Recent advances in ZnO materials and devices, in Material
Science and Engineering, vol. 80, pp. 383-387, March 2001.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
43

[5]

W. Xufeng, B. Hua, L. Chun, L. Gewu, and S. Gaoquan, Controlled onestep fabrication of highly oriented ZnO nanoneedle/nanorods array at
near room temperature, in The Royal Society of Chemistry, pp. 16551657, March 2006.

[6]

Z. Zhongping, Y. Haidong, S. Xiaoqiong, and H. Mingyong, Near-roomtemperature production of Diameter-tunable ZnO nanorod arrays
through natural oxidation of zinc metal, in Chemistry A European
Journal, vol. 11, pp. 3149-3154, 2005.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
44

Wind Turbine Cylinders with Spiral Fins

Radomr Goo, Stanislav Rusek, Miroslav Hrabk

VB-TU Ostrava
17. listopadu 15
Ostrava, Czech Republic

Abstract- Paper describes the worlds first wind turbine

system that rotates with cylinders which have spiral-shaped fins
coiled around instead of common propeller-type blades. When the
spiral cylinders catch the wind, rotating force is generated due to
the aerodynamic properties caused by Magnus Effect. On the
basis of experiments make by MECARO [Link] it is possible to
state that this type of power station is better than propeller type more effective, cheaper the construction cost and stronger in the
strong wind.

Equation (1) demonstrates the manipulation of

characteristics needed to determine the lift force generated by
inducing a mechanical rotation on a ball.

F=

1
V 2 A l
2

(1)

F = lift force
= density of the fluid
V = velocity of the ball
The Magnus effect is the phenomenon whereby a spinning
A = cross-sectional area of ball
object flying in a fluid creates a whirlpool of fluid around itself,
l = lift coefficient
and experiences a force perpendicular to the line of motion and
The lift coefficient l may be determined from graphs of
away from the direction of spin. The overall behavior is similar
experimental
data using Reynolds numbers and spin ratios. The
to that around an aerofoil with a circulation which is generated
spin
ratio
of
the
ball is defined as ((angular velocity * diameter)
by the mechanical rotation, rather than by aerofoil action [1].
/
(
2
*
linear
velocity)).
In many ball sports, the Magnus effect is responsible for the
For a smooth ball with spin ratio of 0.5 to 4.5, typical lift
curved motion of a spinning ball. The effect also affects
coefficients
range from 0.2 to 0.6.
spinning missiles, and is used in some flying machines.
The
lift
coefficient can be approximated using the Thin
German physicist Heinrich Magnus first described the effect
Airfoil
Theory
or Lifting-line theory.
in 1853, but according to James Gleick [2], Isaac Newton
described it and correctly theorized the cause 180 years earlier,
II.
MECARO
after observing tennis players in his Cambridge College.
I.

INTRODUCTION

A. Principle
When a body (such as a sphere or circular cylinder) is
spinning in a fluid, it creates a boundary layer around itself,
and the boundary layer induces a more widespread circular
motion of the fluid. If the body is moving through the fluid
with a velocity V the velocity of the fluid close to the body is a
little greater than V on one side, and a little less than V on the
other. This is because the induced velocity due to the boundary
layer surrounding the spinning body is added to V on one side,
and subtracted from V on the other. In accordance with
Bernoulli's principle, where the velocity is greater the fluid
pressure is less; and where the velocity is less, the fluid
pressure is greater. This pressure gradient results in a net force
on the body, and subsequent motion in a direction
perpendicular to the relative velocity vector (i.e. the velocity of
the body relative to the fluid flow).
B. Calculation of lift force
The KuttaJoukowski theorem relates the lift generated by a
right cylinder to the speed of the cylinder through the fluid, the
density of the fluid, and the circulation.

MECARO Co. was founded in 1998 as MECARO Akita,

and started as the company mainly dealing with designing and
manufacturing machinery used in the production line of
factories.
Wind changes its course so irregularly in Japan that it is said
that Japanese wind pulsates. Akita is a windy region with the
wind blowing from Sea of Japan, and yet almost all the wind
turbines standing here are foreign-made.
With the collaboration of the partnership with the
government, industry and academia, and four years of trial and
error, they were finally able to transform the Magnus theory
into the wind turbine. This was a moment when MECARO
Akita came into being as the wind turbine manufacturer.
MECARO Co., in Akita Prefecture (in northern Japan)
launched large-scale production and sales of the Vortes-model
wind turbine, in April 2007. It is the world's first wind turbine
based on the principle of the Magnus effect, enabling the
turbine to turn using lift generated by spinning cylinders with
spiral fins instead of the usual propeller-like blades.
The wind turbine effort of this company in northern Japan,
where fairly stiff winds often blow, began when the Akita

This work was supported by the Czech Science Foundation (No. GA R 102/09/1842)

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
45

government supplied its founder with information about efforts

in Russia [3] (where turbine power generation has been
ongoing for years) to use the surface effect for power
generation. Mecaro carried out experiments on wind turbine
generators and found a way to enhance the Magnus effect on
the wind turbine "blade" surface by making "striations" on it.
The effect had been identified several centuries ago in
Europe, but Mecaro made an adaptation for wind turbine use
wherein "fins" arrayed into a spiral design were added to
cylinders that have replaced conventional blades. The fivecylindered product markedly improves wind power generation
efficiency even if rotated at a slow pace and furthermore can
withstand gale force weather. Moreover, it is responsive to the
wind's fluctuations while also being quieter than other turbines
available, since the greater lift provided by the fins allows for
lower speed rotation of the propellers.
Theoretical analysis had not been fully developed on the
Spiral Magnus turbine, so the Japanese firm then contacted the
NASA Ames Research Center in California in order to utilize
their giant wind tunnel for testing, which was conducted in
January and February of 2007. The results verified the efficacy
and safety of the system. More improvement work is being
carried out with assistance from Kogakuin University as well
as the University of Tokyo. Mecaro has obtained patents in
Japan and Korea in relation to the Spiral Magnus turbine, in
addition to having other applications pending elsewhere.
According to New Energy and Industrial Technology
Development Organization (NEDO) which partially subsidized
the Mecaro endeavor, there are more than a thousand wind
turbine power generators operating in Japan now, but problems
such as noise pollution and the need to comply with new
building codes to ensure such facilities can withstand typhoons
as well as tremors have cropped up recently.
Considering that other Asian countries prone to typhoons,
cyclones and other natural phenomena would still want to
make use of wind power, it is foreseen that the Mecaro system
will soon gain adherents outside of Japan.
The lift of the Vortes is four times that of conventional
propeller-type wind turbines, and the turbine operates at any
wind speed. Thanks to the greater lift, its rotational speed can
be lowered to around 25 percent of that of propeller-type
turbines, resulting in lower noise levels.
Its cylindrical, durable blades and low rotational speed
reduces this wind turbine's vulnerability to damage caused by
aerial objects. In addition, since the spinning cylinders are
controlled depending on wind speed, the Vortes can operate
without a brake, and there is no concern about excess rotational
speed in the case of strong wind--another safety feature. These
technical improvements make it possible to install the Vortes
even in or near residential areas, where there would otherwise
be noise or safety concerns.
Until now, wind power generation has not been considered a
viable investment for electricity generation except for large
megawatt-scale turbines. In the case of a 10 kilowatt turbine
(equivalent to electrical power consumption of five or six

typical households), however, the Magnus wind turbine has

succeeded in reducing the power-generating cost to 45 yen
(about U.S. 38 cents) per kilowatt, below the cost of solar
power generation.
III. PROPELLER TYPE WIND TURBINE
A common propeller type wind turbine catches wind by the
propeller-shaped blades, and rotate the rotor Fig. 1.
Propellers pick up the wind and generate aerodynamic lift
Fig. 2 (lifting force / rotating force). Fig. 3 shows influence of
Wind Flux Distribution.

Figure 1. A common propeller type wind turbine.

Figure 2. Aerodynamic lift.

Figure 3. Influence of Wind Flux Distribution.

IV. SPIRAL MAGNUS WIND TURBINE

Spiral Magnus is the worlds first wind turbine system that
rotates with cylinders which have spiral-shaped fins coiled
around instead of common propeller-type blades. When the
spiral cylinders catch the wind, rotating force is generated due
to the aerodynamic properties caused by Magnus Effect. With
this principle applied, Spiral Magnus is a product with high
power generation capability and safety.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
46

Each of the five blades (cylinders) of Spiral Magnus Wind

Turbine spins driven by the built-in motor and wind blowing
through the cylinders rotates the rotor (Fig. 4).

Figure 6. Influence of Wind Flux Distribution.

V. HISTORY
The first application was a power ship. 1852 A German
scientist Heinrich Gustav Magnus discovered the phenomenon
of the Magnus Effect. 1926 A German-born aviation engineer
Anton Flettner built a rotor-ship harnessing the power of
Magnus Effect to make the ship move, and made a successful
voyage across the Atlantic (Fig. 7).

Figure 4. Spiral Magnus Wind Turbine.

This system uses a mechanism to generate more lift (lifting

force / rotating force) from the wind power captured by the
cylinders by effectively utilizing the aerodynamic properties of
Magnus Effect (Fig. 5).

Figure 7. Anton Flettners rotor-ship.

1983 The Barrel-blade windmill experiment in U.S. (Fig. 8)

Figure 5. Aerodynamic lift of Magnus Effect.

Each cylinder alone can produce small power, but the five
cylinders working together can generate enough power to
rotate the wind turbine that has a heavy generator in it. The
aerodynamic lift force works in the same way on a propellertype wind turbine and an airplane although the force is created
differently. Fig. 6. shows influence of Wind Flux Distribution.

Figure 8. Hanson's Wind Turbine.

Since then, many countries have tried to develop the

cylinder-blade windmill, but have not been successful to put
into practical use. There was a problem to overcome that the
smooth surface of the cylinders required top speed spinning of

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
47

the cylinders, which consumed more power than the wind

turbine could generate.
MECARO tried to solve this problem by coiling the spiral
fins around the cylinders, and tested at Akita Prefectural
University (Fig. 9).

able to develop an efficient wind turbine that is safe and stable

with low rotational speed, and yet capable of generating high
power. From this point, we set out our long, process of trial and
error and evolution.
VI. APPLICATION
Spiral Magnus is less intimidating because rotational speed
is about a one-sixth of common propeller types. Noise level is
so low that you can hardly hear it through the breeze of natural
wind. There is less chance of bird strike because of the slow
rotational speed.
The rotating cylinders (blades) have excellent power
generating stability since they are automatically optimized to
various wind velocities for efficient wind capturing. The
generating capacity is determined reflecting Japan
characteristic wind pattern of fluctuating air stream and wind
velocities, based on actual field data accumulated through
proof tests at Ogata Village, Akita.
There are two possibilities of connection to system:
1) On-grid System (Interconnection Type) - system
interconnects Commercial Power Supply, working as Sub
Power Supply for Factory or Building
2) Off-grid System (Stand Alone Type) - system is working
as Stand-alone Type Power Supply, or Emergency Power
Supply.
VII.

Figure 9. Spiral Magnus.

Result was that lifting forces gained from the spiral cylinders
was several times larger than those from non-spiral cylinders
(Fig. 10).

CONCLUSION

The experiment plant of 11.5 m in diameter was constructed

in the Akita Prefecture Ogata village Japan for the proof of the
power generation efficiency of a new Magnus Windmill.
There is a big feature that a new Magnus Windmill is better
than past propeller type wind power generation the power
generation efficiency and cheaper the construction cost. A
wind turbine with Magnus effect can be used in a wide range
of wind velocities including storm winds.
Application of the Magnus Effect is a relatively new area
which still leaves plenty of room for analysis to be made. In
parallel with commercializing product, MECARO will
continue to work on the analysis through the joint research
with Tokyo University and Kogakuin University in their
pursuit of product innovation. How can be further improved
efficiency? What shape shall be ultimately perfect? These are
the questions we are seeking answer for.
REFERENCES
[1]
[2]
[3]
[5]

Figure 10. Data per one cylinder.

That is because the spiral fins capture more of wind stream,

and effectively harness the wind as the lifting force. Based on
this finding, there were good possibilities that they might be

[6]
[7]

Clancy, L.J., Aerodynamics, Pitman Publishing Limited, London 1975,

ISBN 0 273 01120 0.
Gleick, James. 2004. Isaac Newton. London: Harper Fourth Estate.
Patent of Russia RU 2189494 C2, 2002.
Shuchi, S., Ito, J., Murakami, N., Kikuchi, H., Kami-yama, S. A study on
Magnus effect for a rotational cyl-inder with an additional structure.
Proceedings of International Conference of Akita Prefectural Univer-sity
2004, Akita, Japan (in Japan).
C. A. Pomeroy, Revolutionary Spiral Fin Turbine, in The Japan
Journal, August 2008 Vol. 5 No. 4.
MECARO [Link]., company web site and documents.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
48

Reliability Standards for Large Interconnections

Petr Horek, Petr Havel

Czech Technical University in Prague, Faculty of Electrical Engineering

Technick 2, 166 27 Prague 6, Czech Republic
{horacek,havelp}@[Link]
Abstract-Every interconnection must be operated so that the
system frequency is maintained at given setpoint and all control
areas of the interconnection meet their active power exchange
obligations. Compliance with reliability standards is required.
The paper discusses performance indices and standards
definitions used in North America and Europe (synchronous
interconnection of 24 European countries joining UCTE - Union
for the Co-ordination of Transmission of Electricity) with the aim
of choosing the most appropriate set. A particular example of
indices and standards applied for the Czech control area of the
UCTE interconnection including their use in Ancillary Services
planning is provided. *

I.

INTRODUCTION

Generation and load in an electric grid must be balanced to

assure stable grid operation. New power market conditions
resulting from the process of liberalization and privatization
dramatically changed the responsibilities of power generation
companies, transmission owners, dealers and other market
players. The process of transfer of ownership and operation of
all high-voltage transmission lines to a regional power pooling
and transmission entity with no interest in generation has been
completed in North America and Europe where linkages
between distribution, transmission, and generation occur across
markets regulated and unregulated rather than through
internal organization. The ultimate instance assuring stable
frequency and defined power flows in the grid are balancing
authorities responsible for individual control areas. Operation
of control areas should be orchestrated so the cooperation, in
terms of exchange of balancing power, is mutually balanced,
otherwise penalties apply.
The state of the power system is described by the frequency
error of the interconnection f and an area control error
ACEarea. Additionally the area net interchange error Parea
might be also considered. The system frequency error is
calculated as
(1)
f = f actual f desired ,
and each control area calculates the Area Control Error (ACE)
as an inadvertent interchange less frequency bias
ACEarea = ( Interchangeactual Interchangescheduled )
+ K area f = Parea + K area f

(2)

This work was supported by the Ministry of Education of the

Czech Republic under the project 1M0567 and by the Czech
TSO under the project Reliability and economy of system
services.

where Karea in MW/Hz is an amount of power that would be

theoretically released through automatic primary control in
generators when the system frequency drops 1Hz from the
scheduled value. Karea is the frequency bias setting for the area.
Interchangeactual is the sum of active power measured on all tielines and Interchangescheduled is the desired, scheduled value.
Interchange is positive when the area is exporting, area
generation exceeds area load. Note that the standard formula
for calculating ACE [1] uses frequency bias coefficient Barea in
MW/0.1Hz (a negative number, not necessarily constant) rather
than K-factor with the obvious relation
(3)
K area = 10 Barea .
B

The load at each time instant is a random variable and this

randomness penetrates to the variables measured and
monitored, such as interconnection frequency and the active
power flows on the tie lines.
These variables are assumed to be random processes
described by the mean value and variance. The correlation
between frequency error and ACE of the area is chosen as a
metric of each areas control performance. We need to impose
an upper bound on that correlation whereas we need not care
about the negative correlation since it indicates a favorable
performance of the particular area.
The generic performance criterion considered says that the
expected correlation between interconnection frequency error
and the areas ACE should be bounded by the allowed
frequency discursion and the required frequency bias of the
area:
2

E {f1 ACE1area } std ( f1 ) ( K area std (f1 ) ) = ( std (f1 ) ) (10 Barea )

(4)
,
where index 1 means a clock-minute (1 minute) average of the
respective variable. The operator E{} is an expected value and
std() stands for standard deviation of the argument. The
principle forms the basis of currently applied standards in
North Americas interconnections and is followed by
Europeans UCTE interconnection in an indirect way as well.
Similarities and differences between performance standards
used in the above mentioned interconnections and a particular
example of standards adopted by the Czech Transmission
System Operator are going to be discussed in the following
chapters.
II. PERFORMANCE STANDARDS

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
49

A. Performance Standards in North America

Today, North American Electric Reliability Corporation
(NERC) defines three Control Performance Standards (CPS)
for the assessment of control areas generation control
performance: CPS1, CPS2 and DCS [1]. All control areas in
North America implemented CPS by 1998.
The motivation to develop the theories that led to the
development of CPS1 and CPS2 came from the need to find
criteria whose fulfillment would equitably divide responsibility
between control areas for satisfying the condition
RMS{fT } = T ,

(5)

where fT is average frequency error over T minutes, T is the

target to be chosen by the interconnection and RMS stays for
root mean square error.
CPS1 requires each Balancing Authority to operate such that,
on a rolling 12-month basis, the scaled average of clock-minute
averages of the ACE of the area multiplied by the
corresponding clock-minute averages of the interconnections
frequency error is less than a specific limit. CPS1 is a statistical
measure of ACE variability. CPS1 measures ACE in
combination with the interconnections frequency error. The
CPS1 requires that the average of the clock-minute averages of
a control areas ACE over a given period divided by its Kfactor times the corresponding clock-minute averages of the
frequency error shall be less than a given constant
ACE1area

f1 12 .
AVG Period
K area

(6)

The constant 1 in Hz is derived from the targeted frequency

bound (the targeted RMS value of one-minute average
frequency error based on frequency performance over a given
year).
CPS1 measures control performance by comparing how well
a control areas ACE performs in conjunction with the
frequency error of the entire interconnection. Criterion (7) can
be viewed as a correlation between ACE and f. Positive
correlation means undesired performance (the area control
error contributes to the frequency deviation from the desired
value) and is therefore limited by the upper bound 12 .
Negative correlation occurs when ACE helps to compensate
the total ACE of the interconnection and is helping to offset the
system frequency deviation.
CPS2 is a statistical measure of ACE magnitude designed to
bound ACE ten-minute averages and provides an oversight
function to limit excessive unscheduled power flows that could
result from large ACE:
AVG10_minute ( ACEarea ) 1.6510 K area Kinterconnection
= L10

(7)

where Kinterconnection is the sum of K-factors in the entire

interconnection and 10 is the targeted RMS of ten-minute
average frequency error. Each Balancing Authority shall
operate such that its average ACE for at least 90% of clock tenminute periods (6 non-overlapping periods per hour) during a

calendar month is within a specific limit, referred to as L10.

which is a statistically derived ten-minute average ACE limit.
Analysis and further discussion on CPS1 and CPS2 can be
found in [5, 6, 7].
The purpose of the Disturbance Control Standard (DCS) is to
ensure the Balancing Authority is able to utilize its contingency
reserve to balance resources and demand and return
interconnection frequency within defined limits following a
reportable disturbance. The application of DCS is limited to the
loss of supply and does not apply to the loss of load. A
disturbance is defined as any event that is 80% of the
magnitude of the control areas most severe single contingency.
A control area is responsible for recovering from a disturbance
within 10 minutes by recovering the amount of the disturbance
or returning ACE to zero. A disturbance is not reportable if it is
greater than the control areas most severe contingency.
Control area must comply with the DCS 100% of the time.
Extra reserves must be carried for the quarter following the
quarter in which the non-compliance occurs.
Each control area can meet the CPS standards by any means
they wish. Some balancing authorities developed AGC logic
that allows meeting CPS and DCS standards automatically.
A control area not meeting the CPS is not allowed to sell
control services to other parties external to its metered
boundaries. This impacts those purchasing control services
from this control area. This is a significant penalty given new
operating environments.
B. Performance Standards in Europe
Speaking about Europe we focus on the largest European
interconnection, the UCTE. According to the UCTE Operation
Handbook (OH) [2] the individual ACEarea needs to be
controlled to zero on a continuous basis in each control area. In
addition, frequency deviation should decay to the given
setpoint in less than 15 minutes and any power outage should
be compensated accordingly. Both large and/or long lasting
ACE deviations should be avoided as much as possible. No
additional explicit requirement on ACE behavior is explicitly
defined. This is because UCTE requires each Balancing
Authority to apply well defined characteristics of the primary
frequency control loops distributed throughout each area and
an area secondary load-frequency controller characteristics are
prescribed as well so the method used by generators and TSOs
is not arbitrary. The only parameter that can influence meeting
the general OH requirements would be the amount of
regulation reserve available for primary and secondary control
and also the level of tertiary reserves available to restore the
capacity of the secondary load-frequency control so it does not
operate close to the saturation level set by the reserved capacity.
In addition, as the tertiary control doesnt run in automatic
mode, the use of tertiary reserve depends on the dispatch rules
of the TSOs operator.
The general quality requests are sometimes specified in more
rigorous way by individual balancing authorities for their
internal use. An example specification is described in the next
section. The mean and the standard deviation of ACE is used

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
50

sometimes for comparing operations of control areas of a

single interconnection [3].
C. Performance Indices Adopted by Czech TSO
Performance indices introduced and used by the Czech TSO
were designed at the time when Czech Republic was part of the
CENTREL control block of the UCTE (CZ,HU,PL,SK).
Internally, there were strict requirements every TSO had to
follow. In addition to the UCTEs OH, |ACE1| should be kept
below 100MW and |ACE60| below 20MW. These requirements
formed the basis for introducing indices for the CZ TSO.
The important principle, which is by the way adopted in
setting limit values for CPS1 and CPS2 criteria, is that the grid
performance should not deteriorate in time. Calculating
individual performance indices on historical data and periods
where the grid operation was generally considered to be
satisfactory, gives the indices setting which we wouldnt like to
exceed in the future. CPS1 and CPS2 set the limit values in
relation to the areas K-factor, which distributes the
responsibility of balancing authorities throughout the
interconnection on fair and well defined basis.
Seven indices describing reliability are defined in [8] and
shown in Table I. Indices rACE1 and rACE60 are statistical
measures of ACE having relation to NERCs CPS2. rACE1t is
linked to UCTE requirement on frequency recovery in less than
15 minutes after a forced generation unit outage and thus it is
related to NERCs DCS index.
While NERCs CPS2 limit L10 gets adapted and may change
from year to year and stationary 90% compliance is required
over years, CEPS limits L1 and L60 are constant and
compliance required is not stationary and gets adapted
according to areas recent performance
TABLE I CZ Performance Indices

rACE1

rACE60

rACE1t

ACE1

MW

ACE1

MW

ACE60

MW

EEvent

MWh

*)

Probability that the absolute average value of oneminute (clock minute) averages of ACE exceeds L1 =
100 MW over given period.
Probability that the absolute average value of onehour averages of ACE exceeds L60 = 20 MW over
given period.
Probability of arriving Events*) over given period.
Average value of one-minute averages of ACE over
given period.
Standard deviation of one minute averages of ACE
over given period.
Standard deviation of one hour averages of ACE
over given period.
Summed energy (absolute value) of all Events
recorded over given period.

The Event shown in Figure 1 is defined as ACE1 exceeding

L1 for at least 10 consecutive minutes. Errors exceeding L1
lasting for less than 10 minutes are not counted. Event energy
is calculated from ACE1 time series.
Event

ACE

Event energy [MWh]

L1=100MW

Figure 1. Event identification.

The limit values for CEPS performance standards (Table II)

are determined statistically from historical records of ACE.
TABLE II Settings from recent ACE evaluation approved by the TSO
Performance index
rACE1
rACE60
rACE1t

Limit value
3.8 %
2.2 %
0.053%

Lets define CZ performance standards in the way reminding

NERCs CPS and DCS standards.
CPSCZ1 is designed to bound ACE one-minute averages and
provides an oversight function to limit excessive unscheduled
power flows that could result from large ACE:
AVG1_minute ( ACEarea ) L1 = 100MW .

Czech TSO shall operate such that its average ACE for at
least (100-rACE1)% of clock one-minute periods (60 nonoverlapping periods per hour) during a calendar year is within a
specific limit, referred to as L1.
CPSCZ2 is bounds sixty-minute averages of ACE and
provides an oversight function to limit excessive unscheduled
power flows that could result from large and/or longer lasting
ACE:
AVG 60_minute ( ACEarea ) L60 = 20MW .

(9)

Czech TSO shall operate such that its average ACE for at
least (100-rACE60)% of clock one-minute periods (60 nonoverlapping periods per hour) during a calendar year is within a
specific limit, referred to as L60.
DCSCZ ensures that the TSO is able to utilize its contingency
reserve to balance resources and demand and return ACE
within defined limits following a reportable disturbance that is
any disturbance resulting in ACE exceeding L1 limit both for
loss of generation as well as loss of load. A control area is
responsible for recovering from a disturbance within 15
minutes in at least (100-rACE1t)% of fifteen-minute periods (4
non-overlapping periods per hour) during a calendar year by
returning ACE below L1.
While control area must comply with NERCs DCS 100% of
the time, the compliance with DCSCZ is weaker. On the other
hand larger number of reportable disturbances are taken into
account because the L1 limit is set much lower.
III. USE OF PERFORMANCE INDICES AND STANDARDS
In order to determine power reserves in the form of AS,
which would technically suffice for proper power balance
control with acceptable reliability, it is necessary to refer to
reliability standards. The principle of using performance
indices rACE1 and rACE60 for determining total reserves
needed to guarantee that the control area will perform
according to the standards is shown in Figure 2 illustrating how
total minimal volumes of power reserves RZ+min 0 and

0
time
10 minutes

(8)

10 minutes

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
51

RZ min 0

are determined from the cumulative distribution

function of ACEOV, estimated ACE of the uncontrolled area.
According to the reliability standards, the absolute value of
ACE is allowed to exceed the threshold 100 MW in rACE1 %
of cases during the defined period. If this splits symmetrically
to positive and negative values, ACEOV higher than 100MW
should be compensated by the control reserves except for (100rACE1/2)% of cases. Hence, we need at least RZ+min 0 and
RZ min 0 reserves to meet the standards of reliability, the
level of which is derived from the control area satisfactory
performance in the past provided by Table II. Reserves
decrease number of occasions when ACE violates the threshold
illustrated in Figure 2 by changing the shape of the probability
distribution function of ACE of the uncontrolled area (curve 1)
to the one of controlled area (curve 2). Legend attached to
point 3 shows how to read the information described by the
probability distribution function associated with ACE.
1 Cumulative distribution function for ACEOV
2 Cumulative distribution function for ACE
%

3 Area Control Error < 200MW found in 87% of all observations (an example)
acceptable number of cases
when ACE1 > limit

100
100 rACE1 2
2
3

not allowed

L10CZ = 1.6510UCTE K CZ KUCTE = 101 MW ,

not allowed

acceptable number of cases

when ACE1 < -limit
reserve

RZ min

reserve

ACE1CZ

f1UCTE = 0.00042Hz 2 < 12UCTE = 0.00044Hz 2

AVG12_month
KCZ

(10)

and the minimum compliance with CPS2, where

[200MW,87%]

rACE1 2
0

ACE as close to zero as possible at all times, as implicitly

demanded by the UCTE Operation Handbook, is more
restrictive as it does not count for solidarity exceeding the level
provided by primary frequency control.
The question what performance standards are better suited
for practical needs, those used in North American or European
interconnection, is difficult to answer as technical and market
conditions differ. There is, however, no much difference in the
performance of the interconnection when looking at the RMS
value of one-minute average frequency error based on
frequency performance over a given year for instance,
1UCTE _ 2007 = 21mHz , 10UCTE _ 2007 = 17mHz , 1Eastern _1998 = 18mHz .
The Czech performance indices were not originally linked to
NERCs CPS1, CPS2 and DCS standards, however a link
between them exists. Thorough investigation is carried out to
find limitations of both sets and decide what indices are more
relevant to practical needs of the ISO/TSO not only for
performance monitoring but also for Ancillary Services
planning.
Note that Czech Area compliance with CPS1 for the year
2007 was over 100% having

ACE1 [MW]

RZ +min

acceptable range of ACE irrespective of number of cases

Figure 2. Cumulative distribution function of the open-loop ACE and its use.

The secondary reserve RZSRmin should compensate for the

fast variations of the open-loop ACE and the tertiary reserves
+

RZTRmin
, RZTRmin
compensate for the slow ACE variations.
The minimal needs of these services are determined similarly
to the total needs but with utilizing cumulative distribution
functions of ACEOV_slow and ACEOV_fast.
The values of the minimal AS requirements are in the form
of time series, which reflects the fact that the behavior of ACE
also varies; typically, it is more uncertain in transitions
periods with changeable weather and temperature, such as in
spring or fall.
More information on the use of reliability standards in power
reserve planning can be found in [4].
IV. CONCLUSION
NERCs CPS1 offers a system wide approach which favors
behavior leading to better overall system performance while
allowing more freedom to individual areas concerning the way
how the areas balancing authority manages its ACE. Keeping

(11)

was 98.2%, which is much better than the required 90% in

DCS. Using (11) for calculating L1CZ and L60CZ provides the
following result
(12)
L1CZ = 122.4 MW, L60CZ = 61.4 MW .
This result indicates that CENTRELs limits L1=100MW and
L60=20MW for CZ were rather strict. On the other hand
compliance with limits (12) would be required in DCS monthly
while compliance with CENTREL limits was required on
yearly basis only.
The paper opens the discussion on selection and use of
performance standards in large interconnections.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]

Reliability Standards for the Bulk Electric Systems of North America,

North American Electric Reliability Corporation, Princeton, NJ (2008)
UCTE Operation Handbook, v2.2/20.07.04, Union for the Co-ordination
of Transmission of Electricity, Brussels (2004)
P. Havel, P. Horek, J. Fantk, and E. Janeek, Criteria for evaluation
of power balance control performance in UCTE transmission grid, Proc.
17th IFAC World Congr., Seoul, 2008
P. Havel, P. Horek, V. ern, and J. Fantk, Optimal Planning of
Ancillary Services for Reliable Power Balance Control, IEEE Trans.
Power Systems, Vol. 23, No. 3, pp. 1375-1382, August 2008
N. Jaleeli, and L.S. VanSlyck, NERCs new control performance
standards, IEEE Trans. Power Systems, Vol. 14, No. 3, pp. 1092-1099,
August 1999
M. Yao, R.R. Shoults, and R. Kelm, AGC Logic Based on NERCs New
Control Performance Standard and Disturbance Control Standard, IEEE
Trans. Power Systems, Vol. 15, No. 2, pp. 852-857, May 2000
G. Gross, and J.W. Lee, Analysis of Load Frequency Control
Performance Assessment Criteria, IEEE Trans. Power Systems, Vol. 16,
No. 3, pp. 520-525, August 2001

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
52

[8]

Rules for the Transmission System Operation Extract from the Grid
Code, Revision 09, EPS a.s., (January 2009) , [online],

[Link]

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
53

Shields of Electromagnetic Wave Based on

Amorphous and Nanocrystalline Soft Magnetic
Materials
Mariusz Ozimek1,2), Dominika Gaworska-Koniarek1), Wiesaw Wilczyski1)
1)

Electrotechnical Institute Wroclaw Division of Electrotechnology and Materials Science (IEL/OW),

55/61 M. Skodowskiej-Curie Street
2)
Wroclaw University of Technology, Faculty of Electrical Engineering,
8 Janiszewskiego Street
Wroclaw, Poland

Abstract- The paper presents results of measurement of

electromagnetic wave shielding effectiveness by shields based on
Fe73,5Cu1Nb3Si13,5B9 soft magnetic material in nanocrystalline and
amorphous form. Manufactured shields were characterised by
good elastic properties. The research was carried out in frequency
band of 300 MHz 1 GHz. The highest shielding effectiveness was
obtained for sample made of amorphous tape - at frequency of
700 MHz it equals 42 dB.

nanocrystalline and amorphous materials. High-tech nano- and

micro- fillers added to polymer warp due to their shielding,
elasticity and easy forming properties can be applied for cables
and wires shielding or as kits protected from EM radiation.

materials
In the experiment the amorphous soft magnetic tape of
Fe73,5Cu1Nb3Si13,5B9
was
used
for preparation of
I.
INTRODUCTION
electromagnetic radiation shields. A part of this tape underwent
Present economic development causes considerable thermal treatment at temperature of 510 C in argon
electromagnetic (EM) radiation emission growth. The source of atmosphere to obtain magnetic nanocrystallites of - -Fe(Si)
this emission are not only TV/radio stations, radar systems but embedded in amorphous matrix. Magnetic properties and
also different kinds of technical devices [1, 2]. Intensive shielding effectiveness were examined for both amorphous and
growth of demand for electric energy, its transmission and nanocrystalline tapes. Magnetic measurements at frequency of
development of electronic, telecommunication and information 50 Hz were performed by means of computerized measuring
technology result in significant increase of electromagnetic system MAG-RJJ-3.0.
Figures 1 2 present the hysteresis loops and total specific
distortion in human environment. In their everyday life people
loss
of examined materials.
incessantly meet with a lot of devices such as microwave ovens,
Test
results indicate that nanocrystalline tape is characterized
screens, inductive heaters, alarms, mobile phones etc. which
by
lower
coercivity (Hc < 1 A/m) and total specific loss than its
emit electromagnetic radiation. The cables supplying electric
amorphous
precursor. Crystallization process resulted also in a
energy for receivers as well as wires used for information
growth
of
magnetic
permeability and saturation induction of
transfer are also sources of EM radiation.
material.
Saturation
induction
of amorphous tape amounts of
To minimize the effect of electromagnetic radiation on living
0.6
T
whereas
this
value
for
nanocrystalline
tape amounts to
organisms usually various types of shields are used. An
about
1.2
T.
Initial
permeability
of
examined
tape
after thermal
electromagnetic radiation shield limits the flow of
treatment
increased
by
about
5
times
and
reached
i100
000.
electromagnetic field between two locations (one with EM
field and the other without it), by separating them with a barrier
II.
EXPERIMENT
made of material with specified electric and magnetic
The EM shields based on amorphous material were prepared
properties. Typically it is applied to enclosures, separating
in
form of a tape with thickness of 20 m and width of 10 mm
electrical devices from the 'outside world', and to cables,
glued
onto thin sheet of paper. The tape strips were arranged in
separating wires from the environment the cable runs through.
such way to eliminate the gaps between the adjacent strips.
The application of shields with proper electric conductivity ,
Nanocrystalline tape due to its high brittleness can not be
permittivity , magnetic permeability and appropriate applied in form of tape strips. Therefore it was pulverized. The
thickness ensures harmonic coexistence of electromagnetic particle size after milling was lower than 80 m. For reference
environment with systems and devices which are the part of purpose the amorphous tape was also milled. Final powders
this environment [3, 4, 5].
dispersed in resin was spread onto textile material and paper
In Electrical Institute Division in Wroclaw research work is sheet. Its mass density on surface amounted to 12 mg/cm2. The
carried out concerning new composite materials designed for sample with addition of carbon fibers dispersed in the same
EM shields. The investigation, among others, is focused on type of resin as in case of basing powder was also
developing of magnetic filler for warp in form of

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
54

Magnetization, Jm (T)

manufactured. The shields obtained in such way were

characterized by good elastic properties.

III. RESULTS

1,2

0,8

0,6

0,4

0,2

0
-200

-150

-100

-50

50

100

150

200

Magnetic field strength, Hm (A/m)

-0,2

-0,4

-0,6
amorphous tape
-0,8

nanocrystalline tape

-1

-1,2

Fig. 1. Hysteresis loop of Fe73,5Cu1Nb3Si13,5B9 alloy in amorphous and

nanocrystalline form
0,30

0,25

Total specific loss, Ps (W/kg)

layer of copper foil (0.1 mm and 0.2 mm). The research was
carried out in frequency band of 300 MHz 1 GHz.

0,20
amorphous tape
nanocrystalline tape
0,15

0,10

Shielding effectiveness of manufactured shields are

presented in Figs. 3 6. The influence of form of applied soft
magnetic material (tape or powder) on shielding effectiveness
is noticeable.
When it comes to shield made of amorphous tape glued onto
sheet of paper the electromagnetic wave shielding changed in
range of 9 dB - 42 dB. The highest one (>40 dB) was noted at
frequency 710 MHz - 740 MHz. Further increase of frequency
resulted in decrease of shielding to 25 dB. In the range of
800 MHz 1000 MHz shielding effectiveness did not
significantly change and amounted about 25 30 dB. It should
be noticed that samples thickness was 0,1 mm. Addition of
thin copper foil did not have a significant effect on shielding
effectiveness results. It indicates that thickness of copper foil is
not sufficient to shield an electromagnetic wave effectively.
Pulverization of amorphous material caused deterioration of
its shielding properties. In this case the shielding effectiveness
is almost constant and does not exceed 10 dB. At frequency of
800 MHz and higher it even dropped below 5 dB. Similar
results were also obtained for milled nanocrystalline tape. The
addition of carbon fibers to nanocrystalline powder brought
about negligible changes in shielding effectiveness.
Shielding effectiveness of specimens made on the basis of
milled nanocrystalline and amorphous soft magnetic materials
is comparable despite the high difference in magnetic
permeability value of blank materials. It might suggest that the
size of particles was not properly matched to frequency band.
Apart from that, the weaker shielding properties of these
samples result from the worse magnetic properties of
pulverized materials. Internal stress induced during milling
process significantly influenced the magnetic permeability
reduction. Additionally non uniform covering of surface by
soft magnetic powder also affects samples shielding
effectiveness.

0,05

10

0,00

0,2

0,4

0,6

0,8

1,2

Magnetization, Jm (T)

Fig. 2. Total specific loss of Fe73,5Cu1Nb3Si13,5B9 alloy in amorphous and

nanocrystalline form (f = 50 Hz)

Shielding effectiveness of examined materials was measured

according to MIL-STD 285 American standard [6].
Transmitting and receiving antenna were situated at a distance
of 3 m from each other in shielded room. Receiving antenna
with wooden support was placed in brass housing which was
padded inside of ferrite layer. A sample was fixed in opening
cut in one of housing side. At first a background measurement
(without the sample) was made. In next step, the examined
samples were placed in housing and their shielding
effectiveness was tested. For shields based on amorphous tape
the tests were conducted for samples with and without thin

Shielding effectiveness (dB)

2
milled nanocrystalline tape with
addition of carbon fibers

0
300

milled nanocrystalline tape

400

500

600

700

800

900

Frequency (MHz)

Fig. 3. Shielding effectiveness of nanocrystalline milled material

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
55

1000

Shielding effectiveness of samples made of soft magnetic

amorphous tape is high what indicates that material in such
form is proper for usage on EM shields
The obtained results of shielding effectiveness in case of
specimens based on pulverized materials are not satisfactory.
These kind of shields require amelioration of preparation
technology and further tests.

50

45

amorphous tape

35

30

25

20

50

15

45

10

40

0
300

400

500

600

700

800

900

1000

Frequency (MHz)

Fig. 4. Shielding effectiveness of amorphous material in tape and powder form

45

Shielding effectiveness (dB)

Shielding effectiveness (dB)

milled amorphous tape

40

35

30

25

20

15
Copper foil of 0.2 mm + amorphous tape
Amorphous tape + copper foil of 0.2 mm

10
40

Shielding effectiveness (dB)

5
35

0
300

400

500

600

700

800

900

1000

Frequency (MHz)
30

Fig. 6. Shielding effectiveness of amorphous tape with copper foil with

thickness 0.2 mm

25

REFERENCES

20

copper foil of 0.1 mm + amorphous tape

[1]

15
amorphous tape + copper foil of 0.1 mm

10
300

400

500

600

700

800

900

1000

[2]

Frequency (MHz)

Fig. 5. Shielding effectiveness of amorphous tape with copper foil with

thickness 0.1 mm

IV.

SUMMARY

The paper presents results of measurement of

electromagnetic wave shielding effectiveness by shields based
on nanocrystalline and amorphous soft magnetic materials. The
research was pilot study with the object of examination of
prospects and usefulness of mentioned above materials as the
fillers of composite electromagnetic wave shields. That is why
the research was concentrated only on pure soft magnetic
materials (or with small addition of carbon fibers) without
composite substrates.

[3]

[4]

[5]

[6]

Sokalski L., Pola elektromagnetyczne i ich niejonizujce oddziaywanie

na organizmy ywe, Przegld Elektrotechniczny, No 4, pp. 100-103,
1999.
Wac-Wodarczyk
A.,
Mazurek
P.A.,
Analiza
zaburze
elektromagnetycznych emitowanych w lokalnej sieci przez komputery
PC, Przegld Elektrotechniczny, No 12, pp. 890 893, 2003.
Sarto M. S., et al, Nanolayered Ligthweight Flexible Shields with
Multidirectional Optical Transparency, IEEE Transactions on
Electromagnetic Compatibility, vol. 47, No. 3, pp. 602-611, 2005.
Lutsev L.V., et al, Dielectric and magnetic losses of microwave
electromagnetic radiation in granular structures with ferromagnetic
nanoparticles. Journa of Physics: Condensed Matter. No 15, pp. 36653681, 2003.
Lundgren U., et al., Shielding Effectiveness Data on Commercial
Thermoplastic Materials, IEEE Transactions on Electromagnetic
Compatibility, vol. 4, pp. 766-773, 2006.
MIL-STD 285, Method of Attenuation Measurement for Enclosures,
Electromagnetic Shielding, for Electronic Test Purposes, U.S.
Government Printing Office, Washington: 1956.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
56

Impedance-Source Inverter-Based High-Power

DC/DC Converter for Fuel Cell Applications
Mikhail Egorov(1), Dmitri Vinnikov(1), Ryszard Strzelecki(2), Marek Adamowicz(3)
(1)

Tallinn University of Technology, Department of Electrical Drives and Power Electronics

Ehitajate tee 5, 19086 Tallinn, Estonia
(2)
Electrotechnical Institute, Poaryskiego 28, 04-703 Warsaw, Poland
(3)
Gdynia Maritime University, Department of Electrical Engineering
Morska 81-87, 81-345 Gdynia, Poland

Abstract- This paper presents the possibility of implementation

of fuel cells with low output voltage range for supplying the highvoltage loads. For matching the different voltage levels and for the
providing of galvanic insulation the isolated DC/DC interface
converter is required. For increasing of power density, efficiency
and flexibility the interface DC/DC converter with three-phase
intermediate AC-link is proposed. The one of features of the
proposed topology is the impedance source inverter utilized in the
input stage of the converter. The paper is devoted to study of
theoretical background of the topology proposed, which is verified
by the simulations.

I.

INTRODUCTION

Fuel cells (FC) have achieved global attention as alternative

power sources due to the environmental concerns. In this paper
the possibility of FC interconnection with the high-voltage load
is studied. Generally, the FCs are used as power sources with
the output voltage variation between 40 V and 80 V. For
supplying the high-voltage, high-power loads it is necessary to
boost the relatively low output voltage of the FC to certain
operating voltage level (for example, 600 V DC), required by
the end user. Moreover, for the safety reasons the isolation
transformer should be used for decoupling the low-voltage
input and the high-voltage output sides of the converter. Thus,
the step-up isolated DC/DC converter topology should be used
in presented application. Practically, it can be realized by
several approaches [1]:
conventional PWM inverter with the step-up transformer
and rectifier,
combination of boost converter with conventional PWM
inverter and step-up transformer with rectifier,
impedance-source inverter with isolation transformer
and rectifier.
The most technically feasible solution is the implementation
of impedance-source inverter (ISI) widely known as Z-source
inverter. The ISI (Fig. 1) employs impedance network coupled
with the main inverter circuit and has unique properties of
voltage step-up and DC/AC conversion [1], [2].
For improving the power density, efficiency and flexibility of
the system the intermediate AC-link of the proposed converter
is realized with the three-phase architecture (Fig. 1). Compared
to recently popular single-phase intermediate AC-link (fullbridge single-phase isolated DC/DC converter) the resulting
advantages of the three-phase approach are obvious:

1. lower RMS current through the inverter and rectifier

switches (higher power transfer through the switch with
the same level of switch current and voltage stresses);
2. reduced isolated transformer's volume (and weight) due to
reduced overall yoke volume and reduced voltage and
magnetic stresses;
3. three-phase transformers and inductors are generally
smaller and lighter than their counterpart single phase
ones, for the same processed power. As a result, the
energy losses in three-phase transformers and inductors
will diminish.
4. reduced size of the output filter due to a dramatic increase
(by a factor of three) of ripple frequency of the dominant
harmonic;
The ISI implemented at the converter input side has the
unique feature that it can boost the output voltage of the fuel
cell by introducing a shoot through operation mode, which is
forbidden in traditional voltage source inverters. Thus, the
varying output voltage of the fuel cell is first preregulated to a
certain value (for example, 600 V) by adjusting the shoot
through duty ratio. Afterwards, the isolation transformers are
being supplied with the voltage with constant amplitude value
and certain duty cycle. With such feature the proposed threephase isolated DC/DC topology with ISI could provide cheaper,
reliable and efficient approach for fuel cell powered systems.
FC

Impedance source inverter

3-phase rectifier

L1

D1

C1

Output filter
Lf

S1

S3

S5

S2

S4

S6

C2

3-phase
transformer

D2

D4

D6

Uin

Cf

Uin1

D3

D5

RL

D7

L2

Figure 1. Proposed topology of the step-up isolated DC/DC converter for

fuel cell applications.

The given paper provides the design guidelines of the proposed

topology. The operation of interface converter will be studied
in two boundary operating points - at the minimum and
maximum output voltages of the fuel cell. The rated power of
the interface converter is 10 kW and the desired output voltage
is 600 V DC. The general specifications of the interface
converter are submitted in Table I.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
57

TABLE I.
GENERAL SPECIFICATIONS OF THE INTERFACE CONVERTER
Input voltage range, Uin
Converter output power, P
Desired DC-link voltage, Uin1
Output voltage, ULoad,
Load current, ILoad
Switching frequency, fsw
Type of switching devices
Peak-to-peak current ripple through Z-source inductors
Desired voltage ripple on the Z-source capacitors

The capacitor voltage during that condition is

40...80 V DC
10 kW
400 V
600 V DC
16.7 A
10 kHz
IGBT
60%
3%

As was reported in previous section, the ISI implemented in

proposed topology has unique features compared with the
traditional voltage source inverter. It is a two port network that
consists of split inductors L1 and L2 and capacitors C1 and C2
connected in X shape (Fig. 1). This impedance network is
basically operating in two states: first is shoot-through when
one or more inverter legs are short circuited. Second state is the
non shoot-through state, when inverter is in active state as a
traditional voltage source inverter. Between the inverter bridge
terminals appears sum of the capacitor and inductor voltage,
referred to DC link voltage in voltage source inverter [1], [2],
[4], [5]. Switching between shoot-through states and non shootthrough states allows boost voltage of capacitors Uc and voltage
of the inverter bridge over input voltage Uin.
During the design of ISI the most challenging is the
estimation of values of the reactive components of the
impedance network [6]. The component values should be
evaluated for the minimum input voltage of the converter,
where the boost factor and the current stresses of the
components become maximal. Calculation of the average
current of an inductor
P .
(1)
IL =
U in
The maximum current through the inductor occurs when the
maximum shoot-through happens, which causes maximum
ripple current. In our design, 60% peak-to-peak current ripple
through the Z-source inductor during maximum power
operation was chosen. Therefore, the allowed ripple current is
IL, and the maximum current through the inductor is ILmax:

I L max = I L + I L 30% .

(2)

I L min = I L I L 30% .

(3)

I L = I L max I L min .

(4)

The boost factor of the input voltage is:

U
1
= in1 ,
1 2 D Z U in

(5)

where D0 is the shoot-through duty cycle:

DZ =

B 1
.
2B

U in + U in1
.
2

(7)

Calculation of required inductance of Z-source inductors:

L=

TZ U C
.
I L

(8)

where T0 - is the shoot-through period per switching cycle:

II. DESIGN RECOMMENDATIONS FOR THE ISI

B=

UC =

(6)

TZ = DZ T .

(9)

Calculation of required capacitance of Z-source capacitors:

C=

I L TZ
.
U C 3%

(10)

Table II shows the initial and calculated values of all the

parameters of the impedance system.
TABLE II.
INITIAL AND CALCULATED VALUES OF THE IMPEDANCE SYSTEM
Parameters
Boost factor B
Shoot-through duty cycle DZ
Z-source inductance L1 =L2, H
Z-source capacitance C1= C2, F
Average current of Z-source inductor IL, A
Maximum current of Z-source inductor ILmax, A
Minimum current of Z-source inductor ILmin, A
Ripple current of Z-source IL, A
Z-source capacitor voltage UC,, V

III.

Fuel cell voltage

40 V
80 V
10
5
0.45
0.4
66
1705
250
125
325
162.5
175
87.5
150
75
220
240

MODELING OF TRANSIENT PROCESSES OF THE

Z-SOURCE INVERTER

To control the ISI all the traditional PWM schemes can be

used. In discussed application to generate a desired output
voltage a modified PWM with shoot-through states is used to
boost the voltage. This control principle of the method is
shown in Fig. 2 (T0 cycle, T cycle of triangular carrier, TZ
time of shoot-through state, UAB,BC,CA line voltages, SA, SB,
SC - signal generators).
The traditional PWM switching sequence based on the
triangular carrier method is shown in Fig. 2. To control the
shoot-through states, the two straight lines are introduced.
When the triangular waveform is greater than the upper
envelope or lower than the bottom envelope, the inverter
switches turn into the shoot-through state. The duty cycle of
the shoot-through state is varying between two predefined
values and is inversely proportional to the fuel cell voltage.
In this circuit topology the parameters of reactive elements,
like L and C are used as obtained from calculations. The
inductors L1 and L2 and capacitors C1 and C2 have the same
inductance and capacitance values, respectively, thus the
impedance network becomes symmetrical [4], [7].

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
58

T0=3T
SA
t

SB
t

SC

Shoot-through

Figure 5. Output voltage and current ripples.

1. Modeling with the fuel cell voltage Uin= 40 V.

T
UAB

Tz/4

Tz/2

Tz/4
t

UBC
t

UCA
t

Figure 2. Control principle of the Z-source inverter.

Simulations were carried out by help of simulation package

PSIM. The simulation circuit is presented in Fig. 3.

Figure 6. Voltage of the Z-source capacitor.

Figure 7. Current of the Z-source inductor.

Figure 3. Simulation model (PSIM) of the proposed converter.

The simulations of the proposed converter were performed

for two operating points with the fuel cell voltages of 40 V
and 80 V, and for the rated output power (10 kW). The
isolation transformer assumed for the simulations was the
delta-delta connected three-phase single core isolation
transformer with the turns ratio of 1:2.5. The simulation results
are presented in Figs. 4 - 13.

Figure 8. Voltage of the primary winding of the isolation transformer.

Figure 4. Voltage of fuel cell and DC-link voltage at the maximum boost factor.

Figure 9. Voltage of the secondary winding of the isolation transformer.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
59

2. Modeling with the fuel cell voltage Uin= 80 V.

Figure 10. Voltage of the Z-source capacitor.

The calculated voltage ripple on the Z-source capacitors is

3% at 40 V and 3% at 80 V (see Table II). The simulated
values of voltage ripple on the Z-source capacitors at input
voltages of 40 V and 80 V were 1.36% and 0.83%,
respectively, which is much better result than the result
obtained by the calculations. The calculated current ripple
through the Z-source inductor is 60% at 40 V and 60% at 80 V.
The values of ripple current received from the simulations were
28% and 57% at input voltages of 40 V and 80 V, respectively.
The output voltage of the DC/DC converter remains constant
(600 V DC) in all boundary operating points. The output
voltage and current ripple were below 1%, which is very good
result for such a powerful system.
V.

Figure 11. Current of the Z-source inductor.

CONCLUSIONS

The paper presented a new step-up isolated DC/DC converter

for the fuel cell applications. By help of theoretical analysis
and simulations it was stated and verified that the
implementation of the Z-source inverter in the primary side of
the isolated DC/DC converter could provide an effective boost
of the input voltage only by the introduction of the shootthrough switching state. The desired DC-link voltage is
preregulated by the variation of the shoot-through duty cycle,
which is inversely proportional to the operating voltage of the
fuel cell.
The simulations showed that the proposed converter is
operating as predicted, without intolerable voltage and current
ripples on its input and output sides. The future work is aimed
to development, assembling and testing of the scaled prototype
(3 kW) of the proposed converter.
ACKNOWLEDGMENT

Figure 12. Voltage of the primary winding of the isolation transformer.

Authors thank Estonian Science Foundation (Grant ETF8020

Research of Advanced Control and Diagnostics Systems for
the High-Power IGBT Converters and Grant ETF7572 High
Power DC Voltage Converters with High Frequency
Transformer Link) for their financial support.
REFERENCES
[1]
[2]
[3]

Figure 13. Voltage of the secondary winding of the isolation transformer.

IV.

ANALYSIS OF SIMULATION RESULTS

The simulations performed show that the proposed ISI is

capable of boosting the varying voltage of the fuel cell to the
desired level (Uin1 = 400 V). After the preregulation for
achieving the required output voltage on the load terminals
(ULoad = 600 V) the further increasing of voltage was realized
by the three-phase step-up transformer.

[4]
[5]
[6]
[7]

R. Strzelecki, and G. Benysek, Power Electronics in Smart Electrical

Energy Networks, 2008 Springer-Verlag London Limited.
Fang Zheng Peng , Z-source inverter, IEEE Transactions on Industry
Applications, Volume 39, Issue 2, Mar/Apr 2003 Page(s): 504 510.
Vinnikov, D.; Isolated DC/DC Converter Topology with a Three-phase
Intermediate AC-Link. 2006 International Baltic Electronics Conference,
2-4 Oct. 2006, pp.1 4.
Fang Zheng Peng and Yi Huang Z-source inverter for power
conditioning and utility interface of renewable energy sources.
J. Rabkowski, The bidirectional Z-source inverter as an energy
storage/grid interface. International conference Eurocon 2007,
Warsaw, September 9-12, pp. 1629 - 1635.
Mitch Olszewski Z-Source Inverter for Fuel Cell Vehicles, Michigan
State University, Department of Electrical and computer Engineering,
Oak Ridge National Laboratory, August 31, 2005.
F. Peng, M. Shen, and Z. Qian, Maximum boost control of the Z-source
inverter, Annual IEEE Power Electronics Specialists Conference, pp.
255-260, 2004, Aachen, Germany.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
60

Voltage division over the interrupter units of a high

voltage circuit breaker without grading capacitors
Henryk Strmer
LS EVH, BTU Cottbus
Cottbus, Germany
Abstract-In this paper the voltage division over the interrupter
units of a high voltage circuit breaker without grading capacitors
is investigated. A method of measuring capacitances and earth
capacitances is presented. High voltage tests were made.

I.

INTRODUCTION

In a modern high-voltage circuit breaker in lower voltage

levels (up to 300kV) one interrupter unit is used per phase. In
higher levels two, three or four switching lines in serial circuit
are used.
Both in off-position and disconnection of the load or short
circuit current the interrupter units are stressed by different
voltages. In order to be able to stress a specified number of
interrupter units by high voltage, the voltage is equally divided
by grading capacitors to all interrupter units. The grading
capacitors are connected in parallel to the interrupter units. For
example for a quad breaking circuit breaker a voltage division
of 25/25/25/25% is aspired and for a twice breaking circuit
breaker a division of 50/50% is aspired. Using available
grading capacitors the voltage division is 53/47% and
27/26/24/23%.
Because the grading capacitors are an important expense
factor for high-voltage circuit breakers for several years its
tried to apply multiple breaking circuit breakers without
grading capacitors. The voltage division is the most important
factor. The result from the calculation of a twice breaking
circuit breaker is 80/20%, for a quad breaking circuit breaker
without grading capacitors it is 80/10/5/5%. In both cases one
interrupter unit has 80% of the voltage; the other units have
only a small voltage.
If you make a dielectric or switching capacity testing you
can see, the circuit breaker is able to keep a higher voltage like
the calculation showed.
The following facts could be the reasons:
- The calculation is too inexact.
- Only the capacities are considered; there could be
More effects.
Its difficult to measure the voltage division between the
interrupting units by common measurement equipment because
the measurement equipment has a noticeable influence to the
voltage division.
The equivalent circuit used consists of capacities. The
capacities are to determine.
The following problems are to research:
- Development of a qualified measuring method:
Measurement of circuit breaker-capacities by

Measurement of capacitive current and with bridge,

Measurement of earth capacities by measurement of
Capacitive current, voltage measurement
Simulation of circuit breaker and measurements:
Field and transient simulations
Several High Voltage Tests: ac-tests and impulse tests
Analysis of the effects, which determine the voltage
Division and which determine the several forms of the
Voltage stress
Analysis of the equivalent circuit
Comparison of the findings and the test readings and
Declare, why the circuit breaker operates better than in
The calculation
Test problems: capacities between test object and
Measuring equipment/environment

II. MEASUREMENT OF CAPACITANCES BY CURRENT

MEASUREMENT

Measurement of the capacitance of a reference capacitor

To determine the capacitance of a capacitor the capacitive
current was measured at two fixed voltages.
A shielded cable was used from the capacitor to the ampere
meter.

Figure 1: Current measurement at a reference capacitor

The test circuit consists of the source, the divider, the

capacitor and the measuring device. The source is a high
voltage transformer and the divider is a capacitive divider. The
test object is a reference capacitor. To measure the current an
ampere meter was used. The measurement was made at two
different voltages. The ohmic resistance and the inductivity are
nearly zero. So the resistance is nearly the reactance of the
circuit. The capacitance can be calculated
u
1
X
(1);
C
(2)
2f X
i
U [kV] i [A]

X [M]

measured Capaci- Capacitance of the

tance C [pF]
reference capacitor [pF]

10,08

328

30,7

103,58

103,63

50,05

1634

30,6

103,92

103,63

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
61

The capacitance calculated from the measured current is

exact. The error is small.
Measurement of the earth capacitance
To measure the full earth capacitance, the ampere meter was
installed on the upper side of the capacitor (on the top). The
ampere meter and the influence to the earth capacitance have to
be small. The ampere meter was located near the capacitor to
minimize the error of the connection from ampere meter to the
capacitor. A short connection had to be used.

Measurement of the current

To determine the capacities of the circuit breaker there were
made several measurements of input and output current. Two
measuring devices were mounted on the left and the right side
of the circuit breaker. The devices were located on a metallic
surface, so the change of capacities is very small.

Figure 5: Current measurement

Figure 2: Measurement of the full earth capacitance

The capacitance was 190pF. A field simulation of the full

earth capacitance follows to a later time.
In the following picture is the first measurement to see. The
ampere meter is on the side of the capacitor, better would be on
top. More tests and simulations will follow to compare several
measure points.

Figure 6: Voltmeter for current measurement

Figure 3: Measurement of the full earth capacitance of the reference capacitor

III. MEASUREMENT OF THE CAPACITANCES OF THE CIRCUIT

BREAKER

The capacities of the twice breaking circuit breaker were

measured by bridge and current measurement. In the following
is an overview of the tests in the high voltage hall. The
measurements were made on three test places, where the circuit
breaker stood.
impulse voltage testing plant
ac voltage testing plant

air conditioning equipment

impulse divider
fence

test place 1

ground

ac divider

amperemeter
test place 3
input
A

circuit breaker
reference
capacitor

amperemeter
output

ground

dc divider

test place 2

dc testing plant

control room
dc divider

The capacities of the circuit breaker are:

- Capacities of the interrupter units C1 , C 2
- Left and right earth capacity C E1 , C E 2
- Middle earth capacity C E
To calculate the searched capacities three equations are
needed.
Many different tests were made:
the output of the circuit breaker opened:
- Interrupter unit 1 and 2 opened
- Interrupter unit 1 and 2 closed
- Interrupter unit 1 closed, interrupter unit 2 opened
- Interrupter unit 1 opened, interrupter unit 2 closed
the output of the circuit breaker grounded:
- Interrupter unit 1 and 2 opened
- Interrupter unit 1 closed, interrupter unit 2 opened
- Interrupter unit 1 opened, interrupter unit 2 closed
The measurements were made up to 50kV, for example in
one configuration the input current was 540A at a voltage of
50kV.
The results of the measurement showed conflicts between
different measurements. It is better to calculate the capacities
only with measurement values from input ampere meter.
If the capacity of the interrupter units is known, for instance
from measurement by bridge or current measurement, there are
only two measurements needed for tests:

Figure 4: test overview

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
62

The input and output cables were long and some meters
parallel to earth. So the operation of the circuit breaker was
imitated and the influence to the capacities was low.
1) Measurement of earth capacities: C = C E + C E1 + C E 2
(3)
Figure 8: An example of measuring by bridge

2) Measurement of interrupter units and middle earth

C = C E + C1 + C 2
capacities:

The biggest capacity was measured on test place 1, the

capacity between source and circuit breaker was the highest
and an error was made.
The source cable was assembled in different places and the
difference between measured capacitances was over 10%.
On test place 2 and 3 the capacity was over 10% smaller than
on test place 1.
(4)

Voltage measurement
The voltage over the interrupter units were measured directly.
The voltage was 20kV. Two voltmeters were used. Because the
input resistance is too low, series capacitances were used.

With the bridge the capacities between the test object and the
source were measured.
Figure 9: Voltage measurement

Picture 7: Example of measurement of capacitance between test object and

source

Some pF were measured and the output current measurement

was wrong because of the leakage capacitance between circuit
breaker and the transformer with the divider.
The second error was the earth capacity.
The measurements were made on three different places for
the test object. If the circuit breaker was near the source the
error of the capacity of the interrupter units was greater. If the
circuit breaker was far away from the source the measured
capacity of the interrupter units was decreased, but because of
the borders in the high voltage hall the earth capacities were
increased. The circuit breaker was here near the side of the
high voltage hall and the control room.
For the first measurements the results show good values. For
a better accuracy the measurements will be repeated in free a
field.

The measuring devices were installed. The first interrupter

unit was shorted and a high voltage was applied. The divider
ratio of measuring device 2 was calculated. The procedure was
made for measuring device 1 with shorted unit 2. When the
ratios were known, the voltages over the units were measured.
Errors are here the leakage capacitance of cables and the
voltmeters with divider capacitances.
Problems of the current and bridge measurements
The measured capacities are very small. Its difficult to
measure these capacities. The cable and the environment have
an influence to the measurement.
These errors have also an influence to high voltage tests on a
circuit breaker.
An example:
An interrupter unit has the capacitance of 15pF and the
middle earth capacity is 40pF. If the circuit breaker is near the
transformer, there are leakage capacities in parallel from the
transformer and divider to the circuit breaker.

Measurement by bridge
The capacity of the interrupter units were measured by a
bridge. The measurements were made up to 50kV.

Figure 10: Leakage capacities of a twice breaking circuit breaker near the
source

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
63

The ac and impulse tests were made with one interrupter unit
and the complete circuit breaker. The 50%-break down
voltages, several statistical parameters and the Weibull
distribution were calculated. The 0,1%- and 99,9%-break down
voltages were calculated from the Weibull distribution.
The voltage distribution over the circuit breaker was
calculated from the ratios of break down voltage of one
interrupter unit over the complete circuit breaker.

grounding

ac voltage testing plant

test place 1

CS12

circuit breaker

CS11

ac divider
CS=CS11+CS12
circuit breaker near the source

Figure 11: Leakage capacities of a twice breaking circuit breaker near the
source

The parallel leakage capacitance CS1 = CS11 + CS12 increases

the capacitance of interrupter unit 1. The voltage ratio is
changed.
An example, where CS1 = 3 pF is.
C is calculated from capacities C2, CE and C1, CS1.
1
U1
C
C=
(5);

100%
(6)
1
1
U C1 + CS1
+
C1 + CS1 C E + C 2
C1 [pF]

CS1 [pF]

CE [pF]

C [pF]

U1 [%]

U1

15

40

11,8

78,6

78,6%

Figure 13: impulse tests

Nr.: 5784
K1
Eval.: LI
Up=
1148kV
T1=
1.0817 s
T2=
51.705 s

Nr.: 57 79
K1
Eval.: LI
Up=
1149.3kV
T1=
1.0938s
Tc=
3.6291s

U2
21,4%

15

40

13,6

75,3

75,3%
24,7%

If the circuit breaker is some meters away from the

transformer and the divider, the leakage capacitance decreases.
IV. HIGH VOLTAGE TESTS
The following high voltage tests were made:
- AC tests with breakdown and withstand alternating
Voltage
- Impulse tests: positive and negative lightning impulse
Voltage and switching surge tests

Figure 14: examples of impulse tests: lightning impulse and chopped lightning
impulse

V. FORECAST
The following points are planned for the future.
- Measurements in free field
- Field simulation of the reference capacitor
- Field simulation of the circuit breaker in the high
Voltage hall
- Transient simulation of the circuit breaker in a network
- Measurement and simulation of a circuit breaker model
- Analysis of the equivalent circuit
- Analysis of the results and the effects of the circuit
Breaker

impulse voltage testing plant

ac voltage testing plant

VI. BIOGRAPHY

air conditioning equipment

impulse divider
Zaun

grounding

ea

rth

ip
str
ing

ac divider

Leiter

circuit breaker

dc divider

dc testing plant

control room
dc divider

Henryk Strmer was born in Cottbus,

Germany. He studied at BTU Cottbus
electrical engineering. He worked several
years in energy distribution in the industry
and in industrial projects at the BTU
Cottbus.

Figure 12: test overview: impulse tests

The ac tests were made up to 700kV and the impulse tests up

to 1150kV.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
64

Adjustable Speed Generation System For Wind Turbine Power Quality

Improvement
Wlodzimierz Koczara koczara@[Link]
Grzegorz Iwanski iwanskig@[Link]
Warsaw University of Technology, Poland
Zdzislaw Chlodnicki [Link]@[Link]
Warsaw University of Technology, Poland &
Cummins Generator Technologies, Stamford, UK
Abstract- The paper presents an emerging technology of electricity
generation based on decoupled theory. The decoupled generation
system operates in wide speed range and control strategy is
oriented to select regions of high efficiency of the driving engine.
This system is used to operate parallel to renewable energy source
that is unstable and not reliable. In case of high power available
from the renewable the decoupled adjustable speed generation
reduces it speed what results in low fuel consumption. However
when the renewable source power is decreasing the adjustable
speed generation system increases its speed following points of
high efficiency of the driving Diesel engine. The paper presents
topologies of the combined systems including variable speed wind
powered generation system and adjustable speed generation
system driven by Diesel engine. Moreover ability of the adjustable
speed to provide power is discussed.

I.

Fig. 1. An example of electricity power demand and renewable available

power

INTRODUCTION

The lack of energy, unstable prices of energy, increasing

prices of fossil fuels and environment protection result in
increase of interest of renewable energy sources. An
alternative to fossil fuel based electricity generation are
usually wind and solar power. However both wind and solar
power depend on weather conditions. Fig. 1 shows an example
of typical variation of power as a function of time. There is a
load power PRL, renewable source available power PRES, and
PD power demanded to keep stability of the system. The
renewable power PRES varies and frequently, in time of high
power demand, this power is not available. Therefore the
renewable source is mainly source of the energy that reduces
general consumption of fossil fuels but not the sources of high
quality power. So to provide high quality power, including
renewable sources, it is needed to provide an additional
generating system that is fully controllable. This additional
generation system will compensate renewable power
fluctuation and will deliver demanded power PD at any time.
Moreover it is advantageous to use controllable energy storage
that acts as instant source of power PSt. Fig. 2 shows topology

Fig.2. Topology of compensated generation system including renewable and

fossil fuel powered electricity sources.

of the generation system that is able to fulfill drawing maximal

energy from renewable source (power PRES). Hence the fossil
fuel based source power PFossil produces high quality power
and usually is much higher than RES power PRES.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
65

The modern variable speed generation systems based on

double fed induction generator (DFIG) are connected to AC
bus as is shown in Fig. 3. A synchronous generator SG
operates with fixed speed according grid frequency.

Power
Grid
PRESS

DFIG
PRESR

RPEC

Fig. 4. History of oil prices.

II. ADJUSTABLE SPEED GENERATION SYSTEM APPLICATION

DF=const

vge

SG

DEF

ge=const

PEV

The adjustable speed generation system [1, 2, 5, 14, 16]

produces AC voltage that frequency is independent to speed.
Such a system (Fig. 5), described as ASGS, is used to

Load
Ex

SDe

EX-S

Fig. 3. Typical connection wind turbine DFIG to AC power grid.

The conventional generation system is driven, for instance, by

an internal combustion engine [15]. Such system efficiency or
specific fuel consumption (g/kWh) depends on load. So, in
case of low power demand i.e. in case high power delivered by
RES, the specific fuel consumption of the engine is high.
Therefore energy saving, responding RES operation, is
accorded by significant losses produced by the engine. It is
then question how economical is the RES operation?
The fossil fuels are strategy issue of supply countries and fossil
fuels prices are not stable and producing great perturbation in
world economy. Fig. 4 shows prices of oil during last 60
years. We remember that in Summer 2008 barrel of oil price
was much higher than $100. Therefore parallel to great efforts,
promoting renewable energy, the same efforts have to be done
in the field of reduction of primary energy consumption by
improvement of efficiency engines fed by fossil fuels.

Fig. 5. Decoupled adjustable speed generation system ASGS as efficient

generation system connected parallel to wind driven DFIG.

compensate the renewable DFIG power fluctuation. The

adjustable speed generation system operates in regions of low
fuel consumption. Fig. 6 shows maximal power of the Diesel
engine Pdmax as function of speed. The adjustable speed

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
66

generation system uses engine power along line A1-A2-A3 i.e.

in the region of low specific fuel consumption g (g/kWh).

Fig. 8. Response of the speed and output voltage UINV of the adjustable
speed generation system on the high step load on the low speed.

To avoid this low quality voltage effect it is used an additional

energy storage system shown in Fig. 9.

Fig. 6. Power and specific fuel consumption of the Diesel engine operating
with fixed and variable speed.

Comparing to conventional fixed speed operation (line C1-C2)

we do notice that most of fixed speed operation is along high
specific fuel consumption.
The adjusted load torque, produced by the generator, is close
to maximum torque. So when series of step load appear then
speed is adjusted and output voltage (Fig. 7) is maintained

Fig. 9. Topology of the connection of decoupled generation system and

variable speed wind powered DFIG.
Fig. 7. Response of the adjustable speed generation system on set of step loads.

(voltage UAB-INV). However when at low speed step of high

power is applied than output voltage is dropping significantly
(Fig. 8)

The energy storage SCE, based on supercapacitor, supplies the

load during short time needed to the engine acceleration. An
effect of the energy storage application is shown in Fig. 10.
The step of load as in case shown in Fig. 10 is not producing

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
67

significant voltage changes because the supercapcitor supplies

converter DC link voltage UDC by the current Isc.

Figure 10. Response of speed , the supercapacitor current ISC and the output
AC voltage UINV on high step load at low speed

III. SUMMARY
The adjustable speed generation performances confirm its
ability to produce power in region of high efficiency. The
system adjusts its power by speed. The maximum power rate is
limited by need of speed acceleration. When an addition
energy storage is applied then system is able to provide high
quality power at any step load. This ability indicates that the
adjustable speed may be selected as future system for
improvement systems including RES especially wind turbine
driven generators.
REFERENCES
[1] [Link], [Link], [Link]-Khayat: Laboratory
Simulation of the Variable Speed Generation System.
EPE Journal Vol. 17 nr 4 Janvier 2008.
[2] R. Srzelecki, G. Benysek: Power Electronics in Smart
Electrical Energy Networks, Springer London, 2008.
[3] Bolognani S., Venturo A., Zigliotto M.: Novel Control
Technique for High-Performance Diesel-Driven
Generator-Sets. Conference Proceedings on Power
Electronics and Variable Speed Drives, 18-19
September 2000. Conference Publication No. 475
IEE 2000, pp. 18-523.
[4] W. Koczara, M. Moskwa, N. L. Brown: Autonomous
Adjustable Speed Decoupled Generation Systems and
their Parallel Operation. Proceedings of 39 Power
Electronics Specialists Conference PESC, 15-19 June
2008, Rhodes, Greece.
[5] W. Koczara, Z. Chlodnicki, [Link], N. L. Brown:
Hybrid Generation System. Proceedings of 3rd
International Conference on Ecological Vehicles and
Renewable Energies, Monte Carlo March 27-30, 2008.
Monaco.

[6] R. H. Staunton, B. Ozpineci, T. J. Theiss, and L. M.

Tolbert, Review of the State-of-the-Art in Power
Electronics Suitable for 10-kW Military Power
Systems, ORNL/TM-2003/209, October 2003.
[7] A. Krasnodebski: Asynchronous Thyristor By-pass
Increasing Short Circuit Capability of Variable Speed
Generating Set. PhD thesis. Warsaw University of
Technology , Warsaw, Poland 2007.
[8] L. Grzesiak, W. Koczara, M. Da Ponte, Novel Hybrid
Load-Adaptive Variable-Speed Generating System,
Proceedings of IEEE Industrial Electronics Conference
Vol.1, p.271-276, Pretoria, 1998.
[9] N. Al.-Khayat, R. Seliga, W. Koczara, B. Kaminski, A.
Krasnodebski, DSP Control of Variable Speed
Integrated Generator, IEEE-ISIE 2002, International
Seminar on Industrial Electronic LAquila, Italy, July
2002.
[10] W. Koczara, L. Grzesiak, M. Da Ponte, Application of
a Permanent Magnet Machine in the Novel Hygen
Adjustable Speed Load Adaptive Electricity Generating
System, IEEE-IEMD Seattle 10-12 May 1999.
[11] G. Iwanski and W. Koczara, Control system of the
variable speed autonomous doubly fed induction
generator Proceedings of International Power
Electronics and Motion Control Conference EPEPEMC04, Riga, Latvia
[12] G. Iwanski and W. Koczara, Sensorless Direct
Voltage Control Method for Stand-Alone Slip-Ring
Induction Generator Proceedings of 11th European
Conference on Power Electronics and Applications
EPE05, Dresden, Germany.
[13] Koczara, E. Ernest, N. Al.-Khayat: VSIG - Variable
Speed Integrated Generation System For Distributed
Generation, Proceedings of the Fifth IASTED
International Conference on Power and Energy
Systems , June 15-17 2005, Benalmadena, Spain, 335340 str., ISBN: 0-88986-465-9, ISSN: 1482-7891.
[14] Hygen variable speed generating system for quality
electric power supply. Technical overview. VoltAmpere, Pretoria, South Africa.
[15] J. Klimstra: Reliable power from renewables with
assistance from reciprocating engines. Managing with
wind Wartisla 2007.
[16] Variable Speed Integrated Generator (VSIG) product
profile. NEWAGE (in present time Cummins
Generator Technologies) Stamford UK.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
68

Current progress in ambient energy harvesting using

piezoelectric materials and electroactive polymers
Pawel Zylka
Wroclaw University of Technology, Institute of Electrical Engineering Fundamentals
Wybrzeze Wyspianskiego 27, Wroclaw, Poland
Abstract- A review on current progress in energy harnessing
micro-generators using active materials (ceramic and polymeric
piezoelectric materials as well as electroactive polymers) is presented.
A problem of effective electrostatic energy extraction from those
capacitive-in-nature energy supplies is illustrated experimentally. It
is shown that using Macro Fiber Composite (MFC) piezoelectric
transducers working in resonant mode and matching the load to the
MFC impedance it is possible to attain 10 % generation efficiency.

I.

INTRODUCTION

Energy harvesting, scavenging, and harnessing these are

all almost analogous terms related to research and engineering
activities aimed at extracting energy in electric form from
various ambient energy reservoirs, which generally cannot be
scaled up for full-size, power-plant energy generation schemes.
There are many ambient energy pools that could be exploited
by autonomous, low-power fuel-less generators: waste heat,
electromagnetic hum, vibrations, localized air movement or
human-generated power. In addition, traditional renewable
energy resources like water flow, tidal and wind energy or sun
radiation can also be exploited at the miniature scale by energy
micro-harvesters. All environments are now being populated
by miniaturized, micro-power electronic devices working in
wireless sensor networks or just as mobile gadgets. A dream to
power all those devices without batteries making them
perpetual or at least to supplement mobile supply scheme is the
driving force of the research oriented on ambient energy
harnessing. Serious interest in those problems, partially
motivated by climate change and global warming, is also
reflected by several R&D organizations (e.g. DARPA, CEALITEN) and companies aiming at harnessing energy from
ambient vibrations, leg and arm motion, shoe impacts, and
blood pressure for self-powered systems.
II. KINETIC ENERGY HARVESTING
Different techniques are being proposed for powering up
remote, unattended, implantable or wearable sensors, shortdistance wireless communication stations, RFID tags or
personal health monitors. Generating electricity from ambient
vibrations or kinetic motion is one of those approaches
applicable as a renewable substitute for batteries in micropower electronic products making timed use of accumulated
energy. Kinetic energy in the form of motion or vibrations is
generally the most versatile and ever-present [1]. Operating
principles of motion-driven ambient energy microscavengers
relay on utilization of inertial forces acting on a proof mass

fitted with a damper which simultaneously serves as an

electromechanical transducer converting kinetic into electrical
energy. Vibration stimuli are extensively present but vary
widely in amplitude and frequency thus micro-generators are
designed as resonant systems matched to vibration spectrum of
the source. Non-resonant or dynamically tunable devices are
also being designed, as broadband operation is required for
maximizing output power when the source has a complex or
time-variable vibration spectrum [2].
A. Electrostatic generators
Although miniature human-powered electromagnetic
dynamos has been used for a long time to supply flashlights
and portable radios yet the electrostatic: capacitive,
piezoelectric and electroactive polymer (EAP) generators
possess a serious advantage over other competing electromagnetic generation schemes: electricity is generated
intrinsically thus virtually all moving parts are eliminated so
the mechanical complexity of the generator is seriously
reduced which makes it lightweight, reliable and wear-proof.
In electrostatic generators, external mechanical force is
engaged to do work against the attraction of charged electrodes
of a variable capacitor. Two modes of such operation are
applicable: switched and continuous. In the switched regime
the capacitive transducer and the supplementary circuitry is
reconfigured by a set of contactors at different stages of the
generation cycle. Such electrostatic transducers require a precharging voltage source in order to operate. As the device
capacitance is reduced against electrostatic attractive forces it
produces current flow into the pre-charging source.
Most of the miniaturized electrostatic energy scavengers are
being build as integrated micromachined chips using MEMS
fabrication processes. Micromechanical vibration energy
scavengers use pre-charged capacitors with capacitancemodulating variable air gap formed between comb electrodes
[3]. In order to avoid pre-charging circuitry electret layer may
be deposited on capacitor electrode - Arakawa et al. has
designed a microseismic electrostatic generator using spincoat-deposited polymeric electret capable of delivering 6 W
of energy for 1 mm amplitude 10 Hz oscillation [4].
Recent years have brought in new concepts emerging on the
rims of high voltage engineering and materials science. These
are related to electroactive polymers (EAP) also described as
artificial muscles. The last name comes from their ability to
directly convert electrical energy into mechanical work and

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
69

motion (linear or bending) without any moving parts found in

classical electrical drives. Electronic EAPs are one of the main
sub-class in broad family of EAP material systems. These are
dielectric polymeric materials that change shape or dimensions
due to shift or limited migration of electrons in response to
electric field stimuli. Dielectric EAP actuators (DEA) display
also inverse effect when working in so-called generator mode
they generate electricity from mechanical deformation. In this
mode DEA can be perceived as a variable capacitance device.
The electrical capacitance is initially charged when the DEA
transducer is in the stretched, high capacitance, state. When it
is then mechanically contracted it goes back to its low
capacitance state and the elastic stresses in the DEA material
film works against the compressive Maxwell stress (electric
field pressure resulting from opposite charges attraction), thus
increasing electrical energy in form of voltage raise. Therefore,
energy harvesting from ambient reciprocating power sources
is also possible by means of dielectric EAP materials.
A recent theoretical analysis indicates that typical humanmotion-generated displacements available for inertial
microgenerators (approx. 5 mm motion at 1 Hz) can provide
energy density of up to 4 W/cm3 and up to 800 W/cm3 can
be extracted from high-pitched machine-induced stimuli
(approx. 2 nm motion at 2.5 kHz) [5]. As typical human
applies over 100 % of his own weight across the shoes at every
heel strike therefore, cushioned shoe soles can be significantly
compressed even during normal walk. Assuming an average
person weighting 75 kg, 5 mm sole compression, regular
walking speed of 2 steps/s this indicates that roughly about
7 W of power could be available just from heel strikes.
Company SRI Intl. has designed and tested a multilayer EAP
shoe generator, schematically illustrated in fig. 1. When the
device is compressed, the coupling liquid or gel medium
contained in the bellows forces out and stretches a thin, elastic
layer of electroded EAP through circular openings in the upper
perforated plate sealing the bellows. In normal walking
conditions such boot energy scavenger was delivering energy
output of 0.8 J per every step with just 3 mm heel compression
and provided 0.8 W of power [6]. The company Artificial
Muscle Inc. has also presented plans to build switched EAPbased power supply units operating on wind or tidal energy.
Flapping (flag- or string-like) power generators made of cyclicdistorted DEA fibers are proposed to populate windy locations
and floats secured to the sea bottom on stretchable elastic
strings made up of dielectric EAP fibers provide extensional
motion by ridding on tidal or wind- generated waves [7].

Figure 1. Schematic view of electroactive polymer heel energy harvester

module developed by SRI Intl. (sketch based on [8]).

SRI Intl. has already started tests of a prototype buoymounted, ocean wave-powered DEA generator. The Hyper
Drive generator utilizes sleeve-like EAP polymer transducers
generating an average output power of more than 5 W under
typical ocean wave conditions, which is enough to power up a
buoy-mounted radio [9].
B. Piezoelectric continuous mode generators
Although theoretical modeling shows that EAP-based
switched energy scavengers using highly compliant
electroactive polymers and running at high voltages can
provide energy density up to 1,5 J/g and much better
performance figures than piezoelectric generators, the latter
(despite their intrinsic material properties limiting efficiency
due to low electromechanical conversion efficiency and high
brittleness and thus low tolerance for high and abrupt
mechanical stress) are extensively utilized in construction of
continuous mode electrostatic generators. The advantage lays
in no requirement for circuitry reconfiguration and no precharging source. In this case mechanical work is partially
stored in the energy of the electric field associated with the
induced polarization of the piezomaterial. Neutralization
current flows if an external load is shorting the electrodes
deposited on the pizoelement. Effective piezoelectric energy
harvesting requires piezoelectric materials with elevated
electromechanical coupling coefficients, thus ceramic (PZT)
materials dominate those designs.
The company Adaptiv Energy is promoting its kinetic
energy harvesting solutions based on Ruggedized Laminate
Piezo (RLP) composite combined with efficient, very-low-loss
MOSFET power extraction electronics. The RLP technology
allows to compress piezoceramic material in order to increase
the voltage per strained unit area. The RLP is a stainless steel
flat spring double-side laminated with a piezoelectric ceramics
and copper electroding layers. Appropriate thermal cycling
during bonding process introduces initial compressive
prestressing into the piezoelectric material, giving greater
density and flexibility and thus provides added output voltage.
RLP transducer uses calibrated mass to match the exploitable
vibration spectrum [10]. Ocean Power Technologies company
exploits strips of piezoelectric foil to harness the power of
flowing water, in the form of 18-inch-long eels that flap in
turbulent water flow like a weeds in the wind [11]. Energy
scavenging by piezoelectric generators has also been recently
perceived by automotive industry, especially car automation
designers and manufacturers. Kim et al. discuss the use of a
piezoelectric cymbal transducers to generate electricity to
supply car remote onboard sensors from the vibration of a car
engine [12]. Dynamic strain coupled from the rolling tire of a
car has also been recently patented as a power source for a
wireless tire condition monitoring system [13].
Energy scavengers operating in continuous mode may also
be used to extract energy from human motion. A heel strike
may be used to flatten a piezoelectric (PZT) unimorph bender
and toe-off action may bend a multilayer polymeric
piezoelectric bimorph stave made of PVDF (polyvinylidene

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
70

fluoride) foil. Such combined system was shown to produce

1.3 to 8.3 mW during a standard walk depending on the
location of transducer (toe/heel) and available displacement
range [14]. The East Japan Railway Company has been testing
an experimental system that produces electricity as people pass
through subway ticket gates. The gates are fitted with floorembedded energy-scavenging mats with series of piezoelement diaphragms that generate electricity as commuters
walk over and compress the mat. 90 m2 energy-scavenging
floor was producing on average 0.14 kWh of energy a day with
peak performance of 0.21 kWh on busy weekends [15].
C. Electric energy extraction from electrostatic scavengers
Another challenging issue is an effective extraction of
electrical energy from capacitive-in-nature, high internal
resistance piezoelectric and EAP elements. Electrostatic
harvester cannot be simply load by ohmic resistor as it results
in poor overall efficiency of the generator. Energy generated
by piezoelectric and most EAP scavengers is available in form
of high voltage peaks which are hazardous for low-voltage
electronics if no appropriate conditioning circuitry is used. To
demonstrate those problems an experiment, described in the
following paragraph, was run using a simple model of the
piezoelectric energy scavenger build of Macro Fiber
Composite (MFC) transducers. MFC is a novel complex
multilayer composite material consisting of PZT piezoceramic
fibers laminated with epoxy resin and sheets of mechanically
tough polyimide foil fitted with interdigitated electrodes.
III.

EXPERIMENTAL

A piezoelectric energy scavenger model was build of two M8557-P1 type MFC transducers (Smart Materials GmbH,
Germany) fixed to a phosphor bronze flat beam (1 mm thick)
using DP460 epoxy adhesive and connected in parallel to
increase magnitude of the generated current. The assembly was
permanently fixed at one end, its free tip was deflected by
3 mm and allowed to freely vibrate at its own resonant
frequency until completely dumped. The voltage produced
over a variable low-inductance load resistor was recorded by
means of a digital oscilloscope fitted with 1:10 10 M passive
probe. An exemplary voltage train recorded for 100 k load
resistance is illustrated in fig. 2. The maximal voltage
amplitude registered in this case was exciding 150 V and even
higher voltages were recorded for high resistive loads (252 V
for 10 M). The signal frequency spectrum contains a strong
component related to the fundamental mode of the beam
mechanical free vibration (approx. 54,5 Hz) as well as a
number of higher harmonics, related to other modes of
oscillations. The recorded voltage trains were then processed to
calculate the mean total energy dissipated at the resistive load the results obtained for different loads are illustrated in fig. 3.
IV.

DISCUSSION

There is an optimal load which results in the maximal energy

output of the MFC energy harvester. The harvester model is
producing oscillatory voltage response, therefore the optimal
load is equivalent to the impedance of MFC transducers which

Figure 2. Exemplary voltage train generated over 100 k resistive load by

MFC energy scavenger model working in resonance mode.

form capacitive-in-nature alternating voltage source. Taking

into account the mechanical resonant frequency of the model
(54,5 Hz) and MFC stack capacitance (21.9 nF, measured for
50 Hz excitation) one obtains impedance of approx. 133 k.
The calculated value is somewhat higher than the optimal load
determined experimentally (approx. 100 k) and the
discrepancy may be attributed to higher frequency components
present in the real voltage response spectrum and to non-flat
frequency-capacitance MFC characteristics. The obtained
results are consistent with findings of Swallow et al. for
resonant harvesters build using piezoelectric fiber composites
(PFC) and polymeric piezoelectrics (PVDF) [16].
The mechanical energy stored in the harvester model beam
displaced by 3 mm equals to 12.9 mJ (the elastic modulus of
the model composite beam has been determined as 2.88 kN/m).
Therefore, even for the optimal electric load, only slightly over
10 % of the stored mechanical energy may be extracted from
oscillating MFC electromechanical transducer model. The
reason for that is a fact that only a portion of the total
mechanical energy pumped into the beam (when it is bent) is
stored in the electric field associated with the piezoelectric
effect due to low electromechanical coupling factor of the
material. Also, some energy fraction is dissipated as the Joule
heat over the MFC internal resistance.
The above simple experiment shows, that it is not only
necessary to transform open-circuit high voltages generated by
piezoelectric transducers into low voltages accepted by
interdigitated circuits. Advanced power converters for use with
piezoelectric and EAP generators are essential in order to
maximize the power harvested as the circuitry needs to present
an optimal load resistance to the piezoelectric/EAP generating

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
71

materials are now the most promising as they allow to put

away most of movable and wearable parts and require very
moderated maintenance efforts. It is expected that the time in
which this technology will be demonstrated for military
applications (so-called smart soldier concept) extends from 2-5
years and 3-7 years is needed in case of domestic appliances
(e.g. mentioned MP3 players) [19].
ACKNOWLEDGMENT
The research work presented herein was financially
supported within the framework of 2007-2009 science funding
program as individual research project no. NN510 2117 33.
REFERENCES
[1]

Figure 3. Energy dissipated at the resistive load by MFC energy scavenger

model working in resonance mode.

element. Most of the recent work on power electronics

intended for kinetic harvesters has been done for piezoelectric
transducers: adaptive control voltage converters, synchronous
half-wave voltage doublers and synchronous rectifiers [17] as
well as new energy storage like supercapacitors or solid state
thin-film batteries has been proposed and tested. Ren at al.
have also shown that combining active energy harvesting
control system (i.e. varying elastic properties of the transducer
between the opposite generation phases by imposing an
external electric field to zero a strain change in the active
material or by imposing other electric boundary conditions to
modify the energy harvesting cycle) together with high
electromechanical yield electrostrictive polymer can provide
harvested electric energy density of approx. 40 mJ/cm3 at
10 % efficiency [18]. Similar approach is also required in case
of EAP harvesters as they are capacitive voltage sources too.
V.

CLOSING REMARKS

MP3 player powered by build-in yoyo energy harvester has

been recently nominated as the best human powered device
design during Core77 Design Challenge 2008 it will provide
round-the-clock listening to music everywhere without a risk
of running out of batteries. And this I just one example of
numerous new gadgets which are just about to come. There is
no doubt that ambient energy harnessing using motion- or
vibration-driven low-power generators is a topic of substantial
and increasing research and engineering attention. Although
there are many different techniques available to harvest raw
energy from ambient energy pools, it seems to be justified to
state that devices using electroactive polymers and piezo-

P. D. Mitcheson et al., Energy Harvesting From Human and Machine

Motion for Wireless Electronic Devices, Proc. IEEE, vol. 96, No. 9, pp.
1457-1486, Sept. 2008.
[2] S. Roundy and Y. Zhang, Toward Self-tuning Adaptive Vibration-based
Micro-generators, Proc. SPIE, vol. 5649, pp. 373384, Feb. 2005.
[3] S. Meninger, J. O. Mur-Miranda, R. Amirtharajah, A. P. Chandrakasan,
and J. H. Lang, Vibration-to-electric energy conversion, IEEE
[Link] Large Scale (VLSI) Syst., vol. 9, pp. 6476, Feb. 2001.
[4] Y. Arakawa, Y. Suzuki, and N. Kasagi, Microseismic power generator
Using Electert Polymer Film, Proc. 4th Intl. Workshop on Micro and
NanoTechnology for Power Generation and Energy Conversion
Applications, November 2004, Kyoto, Japan, pp. 187-190.
[5] P.D. Mitcheson et al., Architectures for Vibration-Driven Micropower
Generators, J. Microelectromech. Sys., vol. 13, no. 3, pp. 429440, 2004.
[6] R.D. Kornbluh et al., Electroelastomers: Applications of Dielectric
Elastomer Transducers for Actuation, Generation, and Smart Structures,
Proc. SPIE, vol. 4698, pp. 254270, 2002.
[7] [Link] link valid on 2009.02.13.
[8] H. Prahlad, R. Kornbluh, R. Pelrine, S. Stanford, J. Eckerle, and S. Oh,
Polymer Power: Dielectric Elastomers and Their Applications in
Distributed Actuation and Power Generation, Proc. ISSS 2005 Intl. Conf.
on Smart Materials Structures and Systems, July 28-30, 2005, Bangalore,
India, pp. SA100-SA107
[9] [Link] link valid on 2007.09.04
[10] P. Buckley, Energy Harvesting Feeds Ultra-low-power Device Growth,
EE Times Deutschland (electronic version available on-line at
[Link] January 2009.
[11] G. Taylor and S. Kammann, Experimental and Theoretical Behavior of
Piezoelectric-Electrostrictive Eel Structures, in DARPA Energy
Harvesting Program Rev., ed. R. Nowak, 2000; ([Link]/dso/).
[12] H. Kim, S. Priya1 and [Link], Modeling of Piezoelectric Energy
Harvesting Using Cymbal Transducers, Jpn. J. Appl. Phys. , vol. 45, pp.
5836-5840, 2006.
[13] Vehicular Mounted Piezoelectric Generator, US patent 4,504,761,
Patent and Trademark Office, 1985.
[14] N.S. Shenck and J.A. Paradiso, Energy Scavenging with Shoe-Mounted
Piezoelectrics, IEEE Micro, vol. 21, no. 3, pp. 3042, 2001.
[15] Y. Takefuji, "Known and Unknown Phenomena of Nonlinear Behaviors
in the Power Harvesting Mat and the Transverse Wave Speaker," Proc.
NOLTA 2008 Intl. Symposium on Nonlinear Theory and its Applications
Budapest, Hungary, Sept. 7-10, pp. 239-244, 2008.
[16] L. M. Swallow, J. K. Luo, E. Siores, I. Patel, and D. Dodds,
A Piezoelectric Fibre Composite Based Energy Harvesting Device for
Potential Wearable Applications, Smart Mater. Struct., vol. 17, pp. 1-7,
2008.
[17] E. Lefeuvre, D. Audigier, and D. Guyomar, Buck-boost Converter for
Sensorless Power Optimization of Piezoelectric Energy Harvester, IEEE
Trans. Power Electron., vol. 22, pp. 20182025, Sep. 2007.
[18] K. Ren, Y. Liu, H. F. Hofmann, and Q. Zhang, An Active Energy
Harvesting Scheme with an Electroactive Polymer, Appl. Phys. Lett., vol.
91, pp. 132910.1-132910.3, 2007.
[19] M. Cain, Characterization of MEMS Energy Harvesting Devices, NPL
Environmental Measures, iss. 3, p. 7, Summer 2007.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
72

Thin Film Adhesion Measuring of Conductive Silicon

Rubber Depending on Power and Coating Time in
Magnetic Sputtering Method
[Link]

Brandenburg Technical University Cottbus

Germany
[Link]
Institute of Electrical Engineering Fundamentals of Wroclaw University of Technology
50-370 Wroclaw, Poland
Abstract- Coating silicon rubber with another material is not
facile task due to the fact, that this substance exhibits very poor
adhesiveness coming from the low surface tension. Adhesion
properties can be improved by means of addition filler and
plasma treatment. The analyzed sample is a conductive silicon
rubber, which was coated using magnetic sputtering method. This
method bases on plasma activity in strong magnetic fields. The
adhesion in modificatied material was evaluated on the basis of
cross-cut test by different coating time and target power.

I.

INTRODUCTION

Silicon rubbers have outstanding properties like excellent

elasticity, temperature-resistant, hydrophobia which are in
great important in high insulation voltage technique. But
silicon rubber can also presents very seriously hurdle in
another possible electrical application due to low surface
tension, which aggravates coating silicon rubber with metal. In
order to increase adhesiveness incorporation of carbon black
and filler can be applied [1].This sample consists of silicon
rubber and carbon black, which after mixing were
manufactured using injection moulding processes. This
material unfortunately is not appropriate for much utilization
e.g. electrodes, high voltage plugs because of insufficient
electrical contact. In order to overcome this problem reliable
coating film is necessary, which due to above mentioned
reason can be achieve only by means advances coating
engineering. Thin film technology, which was used, is called
magnetic sputtering. This method is in generally physical
vapour deposition process. In magnetic sputtering magnetic
field is superposed on the cathode and glow discharge, which is
parallel to the cathode surface. The electrons in the glow
discharge show cycloidal motion, and the centre of the orbit
drifts in the direction of E B with the drift velocity of E/B,
where E and B denote the electric field in the discharge and the
superposed transverse magnetic field, respectively [2].The
construction of magnetron(magnetic) sputtering system is
shown in Fig.1.

Working Gas

subtrate

plasma
Vacuum chamber

Magnetron Power Supply Unit

Pumping
System

Figure1. Magnetron sputtering system

Magnetic Sputtering Method differs from other sputtering

techniques in that the majority of the plasma is confined to
region near the target surface by using strong magnetic fields to
bend the trajectories of the secondary electrons. Advantages,
which conclude: increased deposition rates, reduced substrate
heating during deposition, reduced working gas pressure
requirements had to taken into consideration by choosing this
surface coatings method [3]. Concerning this method there are
many parameters, which have an effect on final coating [4,5].
The aim of this research was to draw a conclusion about
dependency between target power, coating time and
adhesiveness. In terms of power at first we have to distinguish
between target power and circling power, whose are connected
with very important element concerning magnetron sputtering
method - power supply unit. In our experiment magnetron was

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
73

powered with DPS (Dora Power System).This device use a

sinusoidal energy modulation, what enable an achieving great
efficiency and power.
What distinguish this system from the conventional power
supply unit for sputtering application is in spite of mentioned
sinusoidal energy modulation also presence so called circling
power Pc ( see Figure 2) Supply circuit consist of among
others rectifier block, impulse generator and in series
resonating circuit LC. The system of quality stabilization Q
transfers a part of energy back to the capacitor of main supply
system, which as a result appears circling power Pc. If quality
Q is differ from unity than on capacitor C resonating circuit
accumulate surplus energy, which is transferred subsequently
through quality limiter back to power supply unit. That means
that impedance plasma-target is a series silencer resonating
circuit.

TABLE I
PROPERTIES OF SILICON RUBBER

external
diameter
internal
diameter
thickness
Density
Shore A
hardness
Tensile
strength
Elongation at
break
Tear strength

DIN 53 479
DIN 53 505
DIN 53 504
S2
DIN 53 504
S2
AST D 624

mm

152

mm

37

mm
g/cm3

25,4
1,08
40

N/mm2

7,0

650

N/mm

40

In our experiment we have used two methods: scratch test and

cross-cut test
A. Scratch test
DC-M
MAGNETRON

AC-M

The first method, which was used in order to analyze coating,

was scratch test. Scratch test measurement is commonly used
in terms of adhesion evaluation according to scratching
procedure presented in figure 3.

ZII

ZI

Effective PowerPE

Current control

Switchboard unit

rectifier

oscillator

DC

AC
DC

PE
Circling power Pc

Pc

Regulation of target
current

Driver
modulator

Figure2. Dora Power System

II. EXPERIMENTAL DETAILS

Conductive silicon rubbers were supplied by GE Bayer
Silicones. Typical properties of the sample (without coating)
are showed in Tab1. In order to achieve meaningful results 9
samples have been coated, where 4 have been coated with titan
and 5 with aluminium.

Figure3. Principle of scratch-testing: after contact of diamond indenter on the

surface of the coating, the inclined sample is moved with constant speed. The
intender remains fixed and produced an increasing load on the inclined sample
[4]

This method as surmised was not appropriate for our goal due
to influence of substrate material. The influence of the
substrate material means that adhesion test cannot take place
on silicon rubber because the load goes beyond the carrying
capacity of the soft substrate. On another words the coating

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
74

cannot withstand specific for scratch test condition, if the

substrate is not hard enough.
Figure 4 and 5 show the failure of this method- impossibility of
determination critical load for delaminating.

Figure6. Location adhesive tape for cross-cut surface [5]

Figure4. Scratch test faulty results
Penetration depth

60

Residual depth
1

Figure7. Location of cross-cut surface straight before peel off [5]

Figure5. Scratch test faulty results in terms of acoustic emission
Acoustic emission

B. Cross-cut test
The second adhesion test namely cross-cut test was carried
out according to the norm DIN EN ISO 2409. The test consists
in making a series of parallel cuts through the film in one
direction and a second series at right angles to the first. I make
six cuts 2 mm apart in each direction, forming 25 squares
(Fig.6 and 7). Cutting distance was chosen adequate to above
mentioned norm for soft material with layer thickness till 60
m.

Legend
1 adhesive tape
2 coating
3 cuts
4 substrate
a smoothed
b peel off direction

[Link] AND DISCUSSION

As mentioned above scratch test method turned out an
inappropriate method for such king of coated material. The
result of the second method has been juxtaposed in Table 2.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
75

TABLE 2
ADHESION EVLUATION

Lp.

1
2
3
4
5
6
7
8
9

coating
material

target
power

time

pressure

titanium
titanium
titanium
titanium
aluminium
aluminium
aluminium
aluminium
aluminium

kW
2,2
3,1
4,3
5,7
2,0
2,0
2,0
2,0
3,1

min
5
5
6
6
2
2
5
10
5

Tr
9,8 10-3
9,8 10-3
9,8 10-3
9,8 10-3
9,8 10-3
9,8 10-3
9,8 10-3
9,8 10-3
9,8 10-3

adhesion
evaluation
according
to DIN EN
ISO 2409
0
0
0
0
0
0
0
0
0

The evaluation oscillates between 0 till 5, where 5 is the

indicator for the poorest adhesion.
The results pointed clearly out that coating shows excellent
adhesion regardless of target power and coating time and metal
coatings. Figure 6 shows the sample after test with any
indication of peeling off.

substrate surface. The adhesion is better for high values of

kinetic energy of incident species, absorption energy of the
deposit, and initial nucleation density. Effective power PE is
proportional to coating velocity [6], which indicates that with
power increase will also coating thickness rise. In our
experiment the maximal value of target power was limited to
5,7 kW due to considerably sample temperature increase,
which in turn could cause undesirable side effects. The
explanation for outstanding adhesion is twofold. First reason is
that general magnetron sputtering technique guarantee for very
good adhesion. The second one is due to the fact that our
parameters were determinated maybe in too narrow range,
which demand further research.

IV. CONCLUSION
Film adhesion measuring of conductive silicon rubber doesnt
indicate any significant difference in terms of target power and
coating time. Magnetron sputtering coating technique turned
out very good tool in terms of reliable adhesion. Our
experiments show that in spite of poor adhesion properties of
silicon rubber achievement of good coatings are possible.
Unfortunately other properties of this material namely low
harness make some testing method unfeasible.

REFERENCES

[1]
[2]
[3]
[4]

Figure8. View of the sample after cross-cut test

Unfortunately it is very difficult to evaluation adhesion by

means another method due to properties of aluminium. It is
very difficult to solder aluminium, what impede other adhesive
test methods.
The adhesion of a film to the substrate is strongly dependent on
the chemical nature, cleanliness, and microscopic topography

[5]
[6]

E-S. Kim , H-S. Kim, S-H. Jung, J-S. Yoon,Adhesion Properties and
Thermal Degradation of Silicone RubberJournal of Applied Polymer
Science vol.103, 2782-2787, 2007
[Link], M. Kitabatake,[Link] Thin Film Materials Technology,
Springer-Verlag,2004
D.S. Rickerbery, A. Matthews Advanced Surface Coating: a Handbook
of Surface Engineering, Blackie & Son Ltd, 1991
[Link], [Link] , [Link], [Link], Combination of scratchtest and acoustic microscopy imaging for the study of coating
adhesion ,Surface and Coating Technology vol.91,83-90,1996.
Paints and varnishes Cross-cut test (ISO 2409:2007); German version
EN ISO 2409:2007
[Link],Nowoczesne
tchniki
rozpylania
magnetronowego, Electonika,4/2006.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
76

Electrical behaviour of current conducting

silicon rubber
[Link]

Brandenburg Technical University Cottbus

Germany

Abstract- Conductive plastic gain more and more interest in

electrical industry due to few advantages in comparison to metal
like light weight and rustles properties and to some extent lower
material price. The idea of using such materials is not new. These
kinds of materials have been applied for many years in fields like
electromagnetic shielding or voltage grading, where there strongly
entrench his position. Recently new possible applications are
being taking into consideration, which have in common that
conductive plastic should conducting significant among of current.
This paper inspects a mechanism of current conducting be means
of percolation theory and resulting with that consequence.
I.

INTRODUCTION

In order to achieve a very low rsistance in plastic material,

which are at normal state insulator it is necessary to add
conductive fillers, which enhance electrical conductivity. The
most commonly filler used for this purpose is carbon black.
The reason for that are first and foremost the low price, rustproof and opportunity to adjust the carbon black conductivity
depending on combustion process [1]. A mixture of plastic and
high conductive carbon black presents a new material, which
exhibits various value of resistance according to percentage of
conductive filler. The typical characteristic of conductive
plastic is showed in figure 1. The dependency between
resistance and carbon black concentration shows highly
unlined characteristic, which could be elucidated by different
physical processes namely mechanical tunnelling and
percolation.

Figure1. Volume resistivity versus the carbon black content

II.

ELECTRON TRANSPORT IN CONDUCTOR-FILLED PLASTIC

SYSTEMS

A. Quantum Mechanical Tunnelling

If carbon black loading has a very small fraction then the
electron transport is being conducted by means of so called
tunnelling. For the conductive element, the conductance is
merely its ohmic value. But for the insulating silicon rubber,
matters are more complicated. Two conductive particles whose
separation is large compared to atomic dimension see each
other through a resistance controlled by the bulk resistivity of
the silicon rubber itself. However, when the distance is small
( 100 ), electrons may tunnel quantum mechanically
between conductive fillers, leading to a lower resistance than
would be expected from the insulator alone. A comprehensive
examination of normal quantum tunnelling can be found in
Sheng. The basic concept in the Sheng proposal is that there
exists across any tunnel junction a voltage that is sum of two
pieces. One part is simply due to the externally applied source;
the other part is contributes by quantum thermal fluctuations
with the junction barrier [2].
B. Percolation theory
For a small volume fraction of the conducting filler particles,
the resistivity of composite is close to that of silicon rubber
matrix (but not equal to due to above mentioned tunnelling
process).As the volume fraction of the conducting filler
particles increase, the particles come into contact with one
another to form the conduction paths through the composite.
As a result, the resistivity drops by many orders of magnitude
at a critical threshold. One a saturation region of conducting
filler particles is reached; there are a large number of
conduction paths, resulting in a low resistivity [3]. This critical
threshold where the resistivity so rapidly drops is called
percolation threshold. The value of percolation threshold is no
constant but depends of many factors and from one sample to
another fluctuates. The factors, which contribute to this, are:
size, aspect ratio, structure, allocation, roughness of carbon
particles and kind of silicon rubber. Taking into consideration
all this factors and precisely prediction of percolation threshold
is a very difficult task. Generally it can be deduced that a high
aspect ratio and high conductive chemical structure of fillers
contribute to advantage of forming the conducting network, so
that percolation threshold is lower.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
77

III. RESULT OF CURRENT STRESS

A. PTC Effect
The most significant result of current conducting through
conductive silicon rubber was increase of resistance with
temperature rise (see Fig. 2). Because black carbon has
negative temperature coefficient of resistance, that means the
resistant decreases with temperature increase, this effect must
have an explanation in another material properties. A
straightforward explanation is to ascribe the observed
experimental facts to self-heating of sample. The thermal
expansion coefficient of silicone rubber is 2.0-2,5 as big as
other organic rubber ;meanwhile, the thermal conduction
coefficient of silicon rubber is also very high, so, serious self
heating results in increasing temperature of the sample [4].The
thermal expansion coefficient resulting in volume expansion
causes breakdown of conductive pathways what contribute to
resistance increase.

IV. CONCLUSION
Conductive carbon black composites have been researched
in many respects including electrical conduction properties and
dielectric behaviour. Because carbon blacksilicon rubber obey
different conduction mechanisms it is very important to know
accurate and relevant characteristics. Because this material
exhibit positive temperature coefficient the most feasible
application can be associated with current limiter device.

REFERENCES
[1]
[2]
[3]
[4]
[5]

[Link], [Link], Elektrisch leitende Kunststoffe, Carl Hanser Verlag

Mnchen , Wien, 1989
[Link], [Link], [Link], Electron Transport
Processes in Conductor- Filled Polymers , Polymer Engineering and
Science, vol.23, No.1, January 1983
[Link], [Link], [Link], [Link], [Link], , Electrical and elastic
properties of conductor-polymer composites, Journal of Material
Science,34:2979-2985,1999
[Link],[Link],[Link], Resistivity and Thermal Reproducibility of
Carbon Black and Metalic Power Filled Silicone Rubber Heaters, u
Journal of Applied Polymer Science ,vol.94,587-592,2004.
E.S Park,[Link] [Link], Resistivity and Thermal Reproducibility
of Carbon Black and Metallic Power Filled Silicone Rubber Heaters,
Journal of Applied Polymer Science,vol.95,1122-1128,2005.

Figure2. Resistance of conductive silicon rubber as a function of temperature

This effect can be reduced at higher carbon black loading if

the contact pressure will be applied. At higher carbon black
loading, the conductive networks increase and the average
interparticle gap becomes smaller. Therefore, the contact
pressure of particle becomes higher and the network
breakdown process becomes less efficient. As a result, the rate
of increase in resistivity is reduced [5].
B. Switching Effect
During high power switching material absorb high energy,
which leads to partial material ablation. After the high energy
impulse is terminate the material is capable of switching again.
This is in contrast to the switching under low current density
and voltage condition where there are no outward indications
of this switching event.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
78

On Hilbert Transform and its application to

assessment of electrical signal waveform distortion
Tomasz Sikorski

Institute of Power Systems Automation Ltd., Wystawowa 1, 51-618 Wroclaw, Poland

Piotr Ruczewski

Wroclaw University of Technology, Wybrzeze Wyspianskiego 27, 50-370 Wroclaw, Poland

Abstract - In order to achieve full description of the distorted
signals applied method of analysis should take into consideration
two kind of desirable parameters: those which characterizes time
behavior of the waveform, and those which gives information
about frequency components. This paper follows by the idea of
Hilbert Transform (HT) which allows to construct analytic signal
and introduce the term of instantaneous amplitude (IA),
instantaneous phase (IP) and instantaneous frequency (IF).
Unfortunately, application of HT, especially in point of
instantaneous frequency, gives meaningful results for restrictive
class of signals. The accurate representation is achieved for
monocomponents signals. Practical application of HT can be
preprocessed by selected signal decomposition method. This paper
provides some analytic derivation of HT and its digital application
for not decomposed and decomposed signals.

I.

INTRODUCTION

Dynamic complex phenomena require extended performance

of applied signal processing methods. One-dimensional Fourier
spectrum analysis, however very useful and fast digital
applicable, can be insufficient, providing only general
information about extracted signal components, with loss of its
time-varying nature. The frequency analysis can provide the
frequency details, but unfortunately, we dont know when the
frequency changes occurred. In case of dynamic phenomena
the assumption of the stationarity can not be fulfilled. Signal
processing has provided the solution for given problem using
two techniques: joint time-frequency analysis (JTFA) or
decomposition of the signal and calculation of instantaneous
amplitude and instantaneous frequency for particular signal
components. This paper is concentrated on mechanism and
meaning of Hilbert transform (HT) for calculation of
instantaneous magnitude and instantaneous frequency.
Hilbert transform allows constructing orthogonal signal to
the given one. Having the pair of signal and its orthogonal form
we can define complex analytic signal with real part as given
signal and imaginary part as the orthogonal effect of Hilbert
transformation. Obtained analytic signal is complex function of
time. For every time instant we can recalculate amplitude as
well as phase of the local, in point of time, complex value. It
leads to new characteristics, which represents changes of
amplitude and changes of phase in time. Derivative of phase in
time is defined as instantaneous frequency [1],[2],[3],[4].

Successful application of Hilbert transform is possible only

for narrow range of signals. The meaningful results are
achieved in case of monocomponents signals. Thus, in practical
cases, the HT is applied not strictly for investigated signal,
usually containing many components, but for single selected
component, obtained using some signal decomposition method.
There are several signal decomposition methods including
orthogonal polynomial expansion, Fourier expansion or
wavelet expansion. Last work of Huang brought new idea of
signal expansion into the set of so called intrinsic mode
function (IMF), which represents the oscillatory mode
embedded in the signal. The algorithm, which serves the idea
of representation of the signal by the set of IMF oscillatory
modes, is called empirical mode decomposition (EMD) and
was derived by Huang et al. Obtained elements of expansion
can be putted through an examination by Hilbert transform in
order to track instantaneous amplitude and instantaneous
frequency of particular components of the expansion.
Described approach is known as Hilbert Huang Transform
(HHT) [5],[6],[7]. Some selected application of Hilbert
spectrum has found its place also in electrical engineering
[8],[9],[10],[11].
The aim of presented studies is to describe Hilbert transform
technique and to test its usefulness for fundamental benchmark
cases of monocomponent and pseudomonocomponent signals.
In order to investigate the methods several experiments were
performed supported by analytic derivation and digital
simulation. One of the contributions of this paper is the
exploration of Hilbert transform in comparison with its
continuous and digital realization. The intention of presented
results is to familiarize with Hilbert engine as crucial element
in process of calculation of instantaneous amplitude, phase and
frequency.
II. HILBERT TRANSFORM AND INSTANTANEOUS PARAMETERS
The Hilbert transform of given real signal x(t) allows to
obtain its orthogonal form y(t) as follows [1],[2],[3],[4]:
+
1 x ( )
y ( t ) = H { x ( t )} =
d
t
(1)
+
1 y ( )
1
x ( t ) = H { y ( t )} =
d
t

This work is supported by Ministry of Science and Higher Education, Poland, in 2007-2010 as scientific project.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
79

; ( t ) = arg ( z ( t ) )

(3)

Here reveals definition of instantaneous amplitude (IA)

recognized as:
IA( t ) = z ( t ) =

( x (t )) + ( y (t ))
2

(4)

Simultaneously, we can distinguish instantaneous phase (IP)

as phase of analytic signal or as imaginary part of natural
logarithm of complex analytic function:
IP ( t ) = ( t ) = arg ( z ( t ) ) or IP ( t ) = Im {ln z ( t )}
(5)
ln ( z ( t ) ) = ln z ( t ) e j ( t ) = ln z ( t ) + j ( t )

) (

Finally, instantaneous angular frequency can be defined as

derivative of instantaneous phase in time.
d ( t )
(t ) =
dt
z' ( t )
or ( t ) = Im ln ( z ( t ) ) ' = Im
(6)

z ( t )
x ( t ) y' ( t ) y ( t ) x' ( t )
(t ) =
x2 (t ) + y 2 (t )

{(

)}

Additional scaling by factor 1 / 2 leads to instantaneous

frequency (IF):
1
1 d ( t )
IF ( t ) = f ( t ) =
(7)
(t ) =
2
2 dt
Application of Hilbert transform in definition form (1) is
difficult. Effective way utilizes its convolution form (2) and
some relations between Fourier transform and the convolution.
Required in this approach is the Fourier transform of
investigated signal x(t). Having:
F { x ( t )} = X ( j )
j, for > 0
(8)

1
F = j sgn = 0 , for = 0
t
j, for < 0

we can recalculate Fourier transform of orthogonal component

y(t) as:
1
F { y ( t )} = Y ( j ) = F F { x ( t )}
(9)
t
Y ( j ) = j sgn X ( j )

Hilbert transform of given signal can be now uncovered as

inverse Fourier transform:

x ( t ) = 1 sin ( 2 50 t + 0 )

4
3

x
y
IP numerical

a)

2
1
0
-1
-2

t1 =

-3
-4 0

0.01

0.02

0.03

0.04
x
IA

b)
1

0.05
time (s)
100
60

0.06

0.07

3T1
kT1 , k = 0 ,1, 2..
4
0.08

0.09

0.1

IF numerical

c)
50Hz

20
0
-20

IF (Hz)

-60

-1
0

0.02

0.04
0.06
time (s)

0.1 -100 0

0.08

0.02

0.04
0.06
time (s)

0.08

0.1

Fig. 1. Example of numerical calculation of Hilbert transform: (a) investigated

signal x(t) and its orthogonal form y(t) with instantaneous phase IP(t);(b) signal
and instantaneous amplitude IA(t);(c) instantaneous frequency IF(t)
4
3

1 = 0 ; 2 =

a)

2
1

3T1
T
kT1 , k = 0 ,1, 2... ;t 2 = 1 kT1 , k = 0 ,1, 2...
4
2
100
IP1
b)
IP2
60 50Hz

t1 =

IF1
IF2

20
0
-20

IF (Hz)

( x (t )) + ( y (t ))

A. Monocomponent sinusoidal case

The area of interests includes sinusoidal signals, mainly
studied in electrical engineering fundamentals. Thus, firstly we
have
considered
monocomponent
sinusoidal
signal
x ( t ) = A1 sin (1t + 1 ) , supported by analytic derivation
presented in Table I. Fig 1 presents the numerical results when
the investigated signal is 50Hz sinusoidal function with
amplitude equals one and phase initial equals 0. Referring to
the comments in Table I limitation of phase from to
makes instantaneous phase a linear periodical function instead
of linear continuous function. Further numerical calculation of
instantaneous frequency according to (7) brings local delta
Dirac components. Its localization depends on the placement of
artificial phase jump from to , which can be
associated with phase initial of investigated sine signal.
Mentioned influence depicts Fig. 2.
x (p.u.), y (p.u), IP (rad)

z (t ) =

(10)

Construction of complex analytic signal z(t) according to (3)

is the initial point for further calculation of instantaneous
amplitude and frequency. Selected analytic derivation and
computer application of instantaneous amplitude and
instantaneous frequency calculation was performed.

x (p.u.), IA(p.u.)

z ( t ) = x ( t ) + jy ( t ) = z ( t ) e j ( t )

y ( t ) = F 1 {Y ( j )}

IP (rad)

Above definition can be presented in convolution form:

1
y (t ) = x (t )
t
(2)
1
x (t ) = y (t )
t
Having a pair of orthogonal components we can defined
complex analytic signal z(t) as:

-1

-2

-60

-3
-4 0

0.02

0.04
0.06
time (s)

0.08

0.1 -100 0

0.02

0.04
0.06
time (s)

0.08

Fig. 2. Influence of phase initial of sinusoidal component on numerical

application of Hilbert transform: (a) instantaneous phase IP(t); (c)
instantaneous frequency IF(t)

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
80

0.1

TABLE I
ANALYTIC DERIVATION OF INSTANTANEOUS AMPLITUDE AND FREQUENCY OF:

x ( t ) = A1 sin (1t + 1 )

Step 1: Derivation of Hilbert transform of orthogonal component y(t) using Fourier form of given signal x(t) and formulas (9) and (10):
Fourier
x ( t ) = A1 sin (1t + 1 )
X ( j ) = jA1 e j1 ( + 1 ) e j1 ( 1 )
Y ( j ) = j sgn X ( j ) = j sgn jA1 e j1 ( + 1 ) e j 1 ( 1 ) = sgn A1 e j 1 ( + 1 ) e j1 ( 1 )
Inverse Fourier
= A1 e j1 ( + 1 ) + e j1 ( 1 )
y ( t ) = F 1 {Y ( j )} = A1 cos (1t + 1 )

Step 2: Definition of instantaneous amplitude and phase referring to analytic signal z(t):

( A sin ( t + ) ) + ( A cos ( t + ) )
2

IA ( t ) = z ( t ) =

= A1

Comments: Instantaneous amplitude of monocomponent signal confirms constant character of its amplitude.

A cos (1t + 1 )

IP ( t ) = ( t ) = arg ( z ( t ) ) = arctg 1
= arctg ( ctg (1t + 1 ) ) = arctg tg 1t + 1 = 1t + 1
A sin ( t + )
2
2

1
1
1

Comments: Instantaneous phase is a linear function of time with coefficient equals 1 . In practical or digital realization the angle is limited to ,
Introducing mentioned limitation given linear function becomes periodical linear function with period
the function of instantaneous phase obtains the border value

1t + 1

and then jumps to

T1 = 2

. We can distinguish the time constant when

T
3
T
t1 + 1 = t1 = + 1 1 =
1 1 kT1 , k = 0 ,1, 2..
T1

2
2
2
2

Step 3: Calculation of instantaneous angular frequency:

Method 1 using derivative and real and imaginary part of analytic signal:

(t ) =

d ( t )
dt

x ( t ) y ' ( t ) y ( t ) x' ( t )
x2 (t ) + y 2 (t )

==

A1 sin (1t + 1 ) A1 sin (1t + 1 ) 1 + A1 cos (1t + 1 ) 1 A1 cos (1t + 1 )

A12 sin 2 (1t + 1 ) + A12 cos 2 (1t + 1 )

A121
= 1
A12

Method 2 using combined derivative and trigonometry approach:

tg ( t ) =

y (t )

x (t )

cos 2 ( t ) =

cos (1t + 1 )
sin (1t + 1 )

1 + tg ( t )
2

dt
= ctg (1t + 1 )

1 + ctg (1t + 1 )
2

cos 2 ( t )

; sin 2 (1t + 1 ) =

cos ( t )
1

1
t

()
2

1
sin 2 (1t + 1 )
sin (1t + 1 )

(t ) =
1

1 + ctg (1t + 1 )
2

(t ) =

cos ( t )
2

sin (1t + 1 )
2

1 = 1

Method 3 using direct derivative of phase function:

(t ) =

d ( t )
dt

d
1t + 1 = 1
dt
2

Comments: Obtained derivation underline constant character of angular frequency of investigated monocomponent signal. However, taking into consideration
mentioned practical representation of instantaneous phase function as periodical form limited in range
in time instant equals t1 ,we have to consider not only constant function

with artificial phase jump from

but also delta Dirac function in t1.

TABLE II
ANALYTIC DERIVATION OF INSTANTANEOUS AMPLITUDE AND FREQUENCY OF:

x ( t ) = A1 sin (1t + 1 ) + A2 sin (2 t + 2 )

Referring to above steps for monoconponent the instantaneous amplitude, phase and frequency of dual sine signal can be presented as:

z ( t ) = x ( t ) + jy ( t ) = A1 sin (1t + 1 ) + A2 sin (2 t + 2 ) j A1 cos (1t + 1 ) + A2 cos (2 t + 2 )

IA( t ) = z ( t ) =

A12 + A2 2 + 2 A1 A2 cos ( (2 1 ) t + ( 2 2 ) )

y (t )
A cos (1t + 1 ) + A2 cos (2 t + 2 )
IP ( t ) = arctg
= arctg 1
x ( t )
A sin ( t + ) + A sin ( t + )
1
1
1
2
2
2

2
2
1 A1 1 + A2 2 + ( A1 A2 )(2 + 1 ) cos ( (2 1 ) t + ( 2 2 ) )
IF ( t ) =
2
A12 + A2 2 + 2 A1 A2 cos ( (2 1 ) t + ( 2 2 ) )

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
81

to

x (p.u.),y (p.u.),IP(rad)

35

x ( t ) =x1(sin
(50
0 ) t + 0 )
t ) =( 21sin
2 t + 50

a)

30

x
y
IP continuous

25
20
15
10
5
0
-50

0.01

0.02

0.03

0.04

0.06

0.07

0.08

c)

0.09

0.1

IF continuous

50Hz

60

III. CONCLUSION

20
0
-20

IF (Hz)

x (p.u.), IA(p.u.)

100

x
IA

b)
1

0.05
time (s)

-1

Hilbert transform is a one of the tool for creation of analytic

signal which can be treated as initial form of the signal for
calculation of instantaneous characteristic of the signal as
instantaneous amplitude, phase or frequency. Presented in the
paper numerical application confirmes some limitation and
restriction of direct application of the Hilbert method. The
solution is supporting the Hilbert calculation by previous
decomposition of the signal. Application of Hilbert tool, and
calculation of instantaneous parameters for every decomposed
components can bring complex information about time-varying
nature of investigated distortion.

REFERENCES

-60
0

0.02

0.04
0.06
time (s)

0.08

0.1

-100 0

0.02

0.04

0.06
time (s)

0.08

0.1

x (p.u.), x1 (p.u.), x2 (p.u.)

Fig. 3. Example of continuous calculation of Hilbert transform of

monocomponent signal: (a) investigated signal x(t) and its orthogonal form y(t)
with instantaneous phase IP(t);(b) signal and instantaneous amplitude IA(t);(c)
instantaneous frequency IF(t)
a) x ( t ) = 1 sin (100 t ) + 0.3 sin ( 300 t )
x=x1+x2
x1
x2
0.6
1.0

0.02
b)

0.04
0.06
time (s)

0.08

-3

0.1

1.0

0.002

160

d)

140

150Hz

120

0.004 0.006
time (s)

zoom
0.008 0.01

IF continuous
IF1 continuous
IF2 continuous

100

0.5

IF(Hz)

x (p.u.), IA(p.u.)

[1]

-2

-1.0 0

0
x
IA1
IA2
IA

-0.5

-1.0 0

IP continuous
IP1 continuous
IP2 continuous

-1

-0.6

1.5

c)

2
IP(rad)

0.2
0
-0.2

0.02

0.04
0.06
time (s)

0.08

0.1

describing instantaneous amplitude, phase, and frequency are

grouped in Table II. Mentioned instantaneous characteristic of
signal parameters is also shown in Fig. 4, including comparison
with instantaneous characteristics of particular signal
components. We can distinguish general comment that direct
application of Hilbert transform, and its further utilization to
instantaneous parameters calculation, is not sufficient for
distorted signal. The initial decomposition process is required
preceded the Hilbert calculation. Having particular components
of the signal we can perform Hilbert tool and obtain complex
information about its time-varying aspects.

80
60
40

50Hz

20
00

0.02

0.04
0.06
time (s)

0.08

0.1

Fig. 4. Example of continuous calculation of Hilbert transform of dual

component signal: (a) investigated signal x(t) and its components x1(t) and x2(t);
(b) instantaneous amplitude IA(t);(c) instantaneous phase IP(t);(d)
instantaneous frequency IF(t)

True instantaneous amplitude, phase and frequency normally

is not weighted by numerical limitation and can be rather
interpreted as continuous functions. In comparison to Fig 1 we
have constructed Fig. 3, presenting Hilbert transform and
further calculation of instantaneous characteristics of signal
parameters after correction of the phase function in continues
mode, without the limitation , . Presenting trend
represents intuitive expectation for constant amplitude and
frequency of investigated monocomponent sine signal.
B. Dualcomponent sinusoidal case
In order to further exploration of Hilbert idea we have also
reintroduce previous derivation into sum of sinus signals
x ( t ) = A1 sin (1t + 1 ) + A2 sin (2 t + 2 ) . Obtained formulas

Bedrosian E., The Analytic Signal Representation of Modulated

waveform, Proceedings of the IEEE, 1962, pp. 2071-2076.
[2] Bedrosian E., A product theorem for Hilbert transforms, Proceedings of
the IEEE, 1963, pp. 868-869.
[3] Alavi-Sereshki M.M., Prabhakar J.C., A tabulation of Hilbert
Transforms for Electrical Engineering, IEEE Trans. Communications,
vol. Com-20, no. 6, 1972, pp.1194-1198.
[4] Boashash B., Estimating and Interpreting the Instantaneous frequency of
Signal Part 1: Fundamentals and Part 2: Algorithms and Applications ,
Proceedings of the IEEE, vol. 80, no. 4, 1992, pp. 520-568.
[5] Huang N. E. at al, The empirical mode decomposition and the Hilbert
spectrum for nonlinear and non-stationary time series analysis,
Proceedings of The Royal Society 454, 1998, pp. 903-995.
[6] Huang N. E., Shen S. S., Hilbert-Huang transform and its applications,
World Scientific Publishing, 2005.
[7] Rilling G., Flandrin P., One of two frequencies? The Empirical Mode
Decomposition Answers, IEEE Trans. on. Sig. Procc., vol. 56, no.1,
2008, pp.85-95.
[8] Andrade M.A., Messina A.R., Rivera C. A., Olugin D., Identification of
Instantaneous Attibutes of Torsional Shaft Singals using Hilbert
Transform, IEEE Trans. on. Pow. Systems, vol. 19, no.3, 2004, pp.14221429.
[9] Marei M.I, Abdel-Galil T.K., El-Saadany [Link]., Salama M.M.A., Hilbert
Transform Based Control Algotihm of the DG Interface for Voltage
Flicker Mitigation, IEEE Trans. on. Pow. Deliv., vol. 20, no.2, 2005,
pp.1129-1133.
[10] Senroy N., Suryanarayanan S., Riberio P.F., An improved HilbertHuang Transform for Analysis of Time-Varying in Power Quality, IEEE
Trans. on. Pow. Deliv., vol. 22, no.4, 2007, pp.1843-1850.
[11] Sikorski T., Ziaja E., Advanced signal representation methods for
analysis of transient states in embedded generation systems, Systems
Journal of Transdisciplinary Systems Science, vol. 13, 2008, pp. 66-72.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
82

Voltage Dependence of Hologram Generation

Dynamics in Nematic Liquid Crystal Hybrid Panels
Agata Anczykowska, Stanislaw Bartkiewicz, Lech Sznitko, Jaroslaw Mysliwiec
Institute of Physical and Theoretical Chemistry
Wroclaw University of Technology
Wybrzeze Wyspianskiego 27
50-370 Wroclaw, Poland
Abstract - In this paper we present results of the study of
dynamics of diffraction efficiency in hybrid liquid crystal panels
(HLCP) dependent on applied voltage. In the experiment we used
laser to write a hologram by modifying the alignment of liquid
crystal molecules. Simultaneously we used a second laser to read
the recorded gratings. We proposed the model explaining the
dynamics of the gratings formation which is composed of three
different following processes related to the charge carriers
generation and their mobility in the photo conducting polymer.

I.

INTRODUCTION

Nowadays photonic devices with a potential use for

information storage and processing are the subject of
increasing interest and development. The photo induced
refractive index changes in nematic liquid crystals are very
popular among different physical processes enabling the
creation of real time holograms. When properly used in optical
systems they allow for pattern recognition, moving object
extraction, optical filtering, and realization of associative
memories. Implementation of such tasks requires sensitive, fast
and reversible recording media. The process of hologram
formation and erasure generally relies on optical build-up of
transient diffraction gratings in a material. [1-2] These gratings
understood as a spatial modulation of index of refraction or
absorption arise as a result of various non-linear optical
processes which are not well described yet. Many parameters
require optimization, and with some improvements the hybrid
liquid crystal panels (HLCP) could become a very efficient
storage medium for large amounts of data in small areas.
The purpose of the research reported in this contribution
was to investigate the dynamics of diffraction efficiency in
HLCP dependent on applied voltage.
II. EXPERIMENT
A. Sample
The investigated liquid crystal panel was made by
sandwiching nematic liquid crystal between two ITO (Indium
Tin Oxide) glass plates onto which thin layers of photoconducting polymer were spin-coated. Scheme of the
sandwich-type HLCP is shown in Fig.1.
For sample preparation we used poly(N-vinylocarbazole)
sensitized on visible light spectra by doping the
trinitrofluorenone molecules (PVK + TNF).

Figure 1. Hybrid nematic liquid crystal panel (HLCP)

The thickness of this layer was about 100 nm. The

multicomponent nematic liquid crystal mixture E7 (Merck,
Darmstadt, Germany) was used as the optically active
medium.
The distance between ITO plates was set by 10 m spacers.
Surfaces of the photo-conducting thin layers were uniaxially
brushed to preliminary direct the molecules of the liquid crystal
without the influence of an applied voltage.
B. Experimental set-up
All measurements were done in a typical degenerated two
wave mixing experiment set-up which is schematically shown
in Fig. 2 [3].
Nd:YAG laser L1, which operates on wavelength = 532
nm was used to write a hologram by modifying the alignment
of liquid crystal molecules. The writing laser beam was firstly
partially attenuated by the grey filter F and then was divided
into two beams Iw(1) and Iw(2) by the beam splitter BS.
The intensities of these beams were not equal so the gradient
filter GF was used to compensate for the difference.
The mirror M was used to direct beam Iw(2) onto
sample HLCP to allow interference with beam Iw(1)
within the sample volume. Interfering beams caused
modulation of the refractive index n.
A second laser He-Ne L2, which operates on wavelength
= 632.8 nm was used to read the stored image.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
83

where

Det

L1

Ir(0) VR
Ir(+1)
Sh
Iw(1)
HLCP

Ir

If absorption of the refractive material is insignificantly

small, the Bessel function parameter for thin diffraction
gratings can be written in the form of
2nd
(5)
=

F
BS

Iw(2)

d the thickness of the liquid crystal panel

wavelength of the writing laser
modulation of absorption coefficient
n modulation of refractive index

GF

For the writing laser wavelength of light of = 632.8 nm

and the thickness of the liquid crystal panel, d = 10m
L2
Os
M.
parameter value 2d/ = 99.1.
The sinusoidal light modulation caused by interfering light
in the volume of sample generates charge carriers in PVK:TNF
Figure 2. Experimental set-up for degenerated two wave mixing measurements.
layers. Number of generated carriers is proportional to the
L1 Nd:YAG laser, L2 He-Ne laser, F grey filter, GF gradient filter, Sh
intensity of incident light, so generation is strongest in the
shutter, BS beam splitter, M mirror, S sample, VR voltage regulator,
Det detector, Os - oscilloscope.
places where constructive interference occurs and none where
destructive interference occurs.
Power of the first order diffraction beam Ir(+1) was
When the sample is illuminated, DC voltage is applied and
measured by silicon detector Det coupled with digital when its value is over Freedericksz potential, the generation of
oscilloscope Os.
the charges and their drift into the photo-conducting layers
Power Supply "VR" of DC voltage connected to the sample provides changes to the local electric field, which results in
was regulated within the range of 0-50V. Diffraction efficiency liquid crystal molecules reorientation and then spatial
was measured in dependency of applied to the HLCP voltage. refractive index modulation in the sample.
All the following parameters were fixed. The reading angle
The charge generation and transport in HLCP can be
(angle between normal to the sample surface and reading beam) divided into three different following processes (n1, n2, n3).
was set to = 12, the writing angle (the angle between writing These three processes are characterized by the following
beams Iw(1) and Iw(2)) was set to 2= 4 and the tilt angle (the parameters:
angle between bisector of writing beams and normal to the - amplitudes determining the quantitative contribution of each
sample surface) was = 0. The power of writing beams were of the processes. (n1M, n2M, n3M)
Iw(1) = Iw(2) = 19mW and the power of reading beam was Ir = - stable in time angles between different diffraction gratings
15mW.
(1/2, 1/3, 2/3) determining their relative position
- time constants (1, 2, 3) determining the rate of exponential
growth
of gratings writing
III. THEORY
- time constants of erasing functions (1e, 2e, 3e)
Experimentally diffraction efficiency is calculated from
Basing on the preceding assumptions we developed the
the measured intensity of light in a certain diffraction order. following mathematical model for description of modulation of
This figure is determined as the ratio of intensity of diffracted the refractive index during diffraction gratings writing process
light in a first diffraction order Ir(+1) to the intensity of which is based on gratings coupling [6].
incident light Ir.
n = [n12+n22+n32 + 2n1n2cos(1/2) +
Ir (+1)
(1)
=
2n1n3cos(1/3) + 2n2n3 cos(2/3)] 1/2
(6)
Ir
Theoretically diffraction efficiency (for thin diffraction
gratings) is determined by using the Raman-Nath
(7)
ni = ni M (1 e t / i ) e t / ie , i = 1, 2, 3
approximation. This approximation can be used to link
diffraction efficiency with the alteration of the composite
IV. RESULTS AND DISCUSSION
refractive index by n* Bessel functions J1 [4,5]

= J 1 ( )

2n * d

n* = n + i

(2)
(3)
(4)

Temporal evolution of diffraction efficiency was measured

for different values of applied to the HLCP voltage. The
experimental data indicate that diffraction efficiency generally
increases with increasing voltage. This is due to the fact that
the number of generated charge carriers responsible for the
efficiency of the process of hologram writing depends on
applied to the HLCP voltage. Temporal evolution of diffraction
efficiency for different values of applied to the HLCP voltage
is shown in Fig. 3.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
84

TABLE I
THE SET OF APPROXIMATION PARAMETERS FOR THE SERIES OF
EXPERIMENTAL CURVES OF WRITING DYNAMICS OF THE HOLOGRAM
CALCULATED USING THE EQUATION (6)

0.020
0.016
0.012

n1M

n2M

n3M

[V]

[ms]

[ms]

[ms]

[10-5]

[10-5]

[10-5]

32

20.0

56.0

580.0

0.0

26.0

37.0

36

20.0

47.0

554.0

70.8

104.9

98.7

0.008

40

20.0

39.5

534.0

97.2

125.9

113.0

0.004

44

20.0

33.5

517.0

154.6

182.9

170.7

48

20.0

28.0

504.0

237.1

259.5

264.4

50

20.0

25.5

498.0

294.8

306.7

288.0

0.000
0

200

400

600

800

1000

time (ms)

Figure 3. Temporal evolution of diffraction efficiency for different values of

applied to the HLCP voltage.
Black curve 32V, dark grey curve 40V, light grey curve 50V.

Diffraction efficiency depends on the modulation of the

effective refractive index, which described by formula (6)
consists of three different following processes (n1, n2, n3).
For the processes of diffraction gratings formation most of the
parameters characterizing those three processes change when
applied to the HLCP voltage changes as shown in Table 1.
The first process which occurs in first few milliseconds
after irradiation of the sample can be explained as generation of
electrons and holes in the bright regions of both PVK:TNF
layers, diffusion of holes into the dark regions of polymer
layers on anode side (redistribution of holes near the border
between PVK:TNF layer and liquid crystal at anode side).
Time constant of the first diffraction grating (1 = 20ms)
remains constant, so it can be concluded that it is independent
of voltage applied to the HLCP. Amplitude of this process (n1M)
increases with increasing voltage and time constant of the
erasing function (1e) decreases.
The next process occurs within several tens of milliseconds
and is connected with the depletion of faster holes at cathode
side and charge transfer in external electric field. Time constant
of this process (2) changes linearly with changing inverse
voltage and time constant of the erasing function (2e) has
infinite value so it can be assumed that it has no effect on the
process of writing of the second grating.
The last process takes a few hundred milliseconds. Charge
is transferred from side holes located on anode to the molecules
of liquid crystal which results in the flow of the cations through
the sample volume and their arrival to PVK:TNF layer at the
cathode side. This process leads to neutralization of electrons
present in the cathode. Time constant of this process (3) also
changes linearly with changing inverse voltage and time
constant of the erasing function (3e) has infinite value similarly
to the second process so it can be assumed that it has no effect
on the process of writing of the third grating.

1e

2e

3e

1/2

1/3

2/3

[V]

[ms]

[ms]

[ms]

[rad]

[rad]

[rad]

32

1300

2.81

0.00

2.81

36

1050

3.01

-0.01

3.02

40

858

3.00

-0.05

3.05

44

700

3.04

-0.06

3.10

48

580

3.05

-0.08

3.13

50

520

3.06

-0.09

3.15

Approximated curves very accurately coincide with the

experimental data obtained for various experimental conditions.
In Fig.4. experimental data of dynamics of diffraction
efficiency and estimated values for the process of hologram
writing for applied to the HLCP voltage U = 40V are shown.
After shutter release a rapid increase in the diffraction
efficiency was observed. After 25 ms it reached the local
maximum value of 0.0067%, then started to decrease and
reached a minimum value of 0.0043% in the 55 ms of
experiment, then rise again to the maximum value of 0.0104%
in the 165 ms of experiment and decreased slowly to the value
of 0.0044% and stabilized after 1000 ms from the start of the
experiment.
0.012
diffraction efficiency (%)

diffraction efficiency (%)

0.024

0.010
0.008
0.006
0.004
0.002
0.000
0

200

400

600

800

1000

time (ms)

Figure 4. Typical time dependency of diffraction efficiency for HLCP.

Gray points experimental data, black points estimated values (parameters of
estimation can be found in Table 1.)

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
85

V. CONCLUSION
In conclusion, we have demonstrated the results of the
comparison between experimental and estimated data.
The presented mathematical model which takes into
account the existence of charge carriers (holes and electrons)
and their mobility in the photo conducting polymer fairly well
describes the three-step process occurring inside the HLCP
during the hologram writing.
The calculated parameters characterizing proposed
mathematical model show that the dynamics of the hologram
writing is linked to the generation and transport of charge
carriers (diffusion and drift) in the external electric field.
ACKNOWLEDGMENT
The authors wish to thank the Wroclaw University of
Technology for financial support.
REFERENCES
[1]
[2]
[3]

[4]
[5]

[6]

I.C. Khoo, Liquid Crystals Physical Properties and Nonlinear Optical

Phenomena, [Link], New York, 1995.
S. Bartkiewicz, A. Miniewicz, A. Januszko, J. Parka., Dye-Doped
Liquid Crystal Composite for Real Time Holography, Pure and
Applied Optics, Vol. 5, p. 799, 1996.
J. Mysliwiec, D. Jarzab, K. Janus, [Link], Study of SelfDiffraction Process in Photoconducting Polymer-Nematic Liquid Crystal
Hybrid Structure Applied Physics Letters, Vol. 90, p. 121120, 2007.
H.J Eichler, P. Gunter, D.W. Pohl, Laser Induced Dynamic Gratings
Springer - Verlag, Berlin, 1986.
A. Sobolewska, S. Bartkiewicz, Three Gratings Coupling During the
Holographic Grating Recording Process in Azobenzene-functionized
Polymer, Applied Physics Letters, Vol 92, p. 253305, 2008.
N. Reinke, [Link], T. Fuhrmann, H. Franke, R.A Lessard, Electric
Field Assisted Holographic Recording of Surfac Relief Gratings in an
Azo-glass, Aplied Physics B, Vol. 78, pp.205-209, 2004.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
86

Relevance of a Safety Factor for Wind Power Trading

in Comparison with the Utilization of a Storage
S. Vller, J.F. Verstege

Institute for Power System Engineering, University of Wuppertal

Rainer-Gruenter-Strae 21
42119 Wuppertal, Germany

This work analyses the opportunities for the participation at

an energy exchange and at the control power market. To interact
with the market, wind farms have to supply a constant power for
a specified time. Due to the fluctuations of wind energy it is not
possibly to guarantee a constant power based on wind forecasts.
Thus, the wind farm cannot sell their maximal estimated energy
but only a reduced value. The aim of the work is to figure out this
specific "safety factor". Later on, the results are compared both
technically and economically with the use of a storage unit, where
the full amount of energy can be traded.

I.

INTRODUCTION

Renewable energy is a constant growing part of the

worldwide energy supply. In Germany the most important
source is wind energy. In comparison to conventional power
plants the most renewables have the main disadvantage of
delivering fluctuating energy why they cannot be used to
guarantee a steady supply.
In spite of this disadvantage, the renewable energy is extra
supported because it is a sustainable and environment friendly
energy form and an important alternative for future supply. In
Germany this official support is regulated by law by the
government through the Renewable Energy Sources Act (EEG)
[1]. The owners of renewable production facilities get a priority
feed-in into the grid, where the grid operator must take-over
the energy and pay a fix refund for the generated energy. This
payment is often higher than the average production cost in
conventional power plants but sometimes lower than
achievable prices at energy markets.
II. SAFETY FACTOR
To take part at an energy market the production facility has
to provide a constant power for a specified time (e.g. 15
minutes or 1 hour). This is the so called "scheduled energy".
Fluctuating energy sources like wind energy have not the
ability to supply this scheduled energy because the forecast of
the next day wind speed is not so accurate. The prediction of
the produced energy is possible, but during the day the wind
power may be not in the time when it is awaited. Due to wind

delay the expected power may shift by one hour. Because of

this forecast uncertainty, the wind farm owner cannot trade the
whole produced wind energy. He has to insert a safety factor,
by which the maximal possible wind power is reduced. By
reducing the power, the wind farm owner is able to supply a
(lower) scheduled power, but will not violate his supply
contract in the case, that the wind power is another than
predicted.
An example is shown in Fig. 1. Concerning to the wind
forecast (blue line) the wind farm owner sells the wind energy
in 1-hour-contracts (orange + grey bars) to a costumer. The
height of power he uses is always the maximal possible power
for this one hour, what is also the minimal wind power in this
time. At the day of delivery the supplied wind power has
another curve (red line), so a part of the sold energy can not be
supplied and the before sold scheduled energy have to be
reduced (grey bars). Thus, the wind farm owner has to pay a
penalty because he cannot fulfill his contract.
Traded Energy
Energy Deficit
Wind Power Forecast
Wind Power real

Energy, Power

Abstract- In Germany renewable energy sources are paid by

law with a constant amount for the supplied energy. Thus, the
operators of such plants have no intention to take part at the
energy market yet, but there exists possibilities to earn a better
payment, so already some facilities act in the market and in the
future this participation will be more and also necessary.

9 10

11 12 13 14 15 16 17 18 19 20 21 22 23 24

Time (h)

Figure 1. Example of energy deficit by trading the maximum power

depending on the forecast compared to the real wind power

By this situation, the wind farm owner has to decide,

whether it is cheaper for him to pay the penalty and sell the
maximal possible scheduled energy or he reduces the
scheduled energy by a safety factor (e.g. he sells only 80 % of
the maximum). Doing the last option, he loses income because
he cannot sell this energy to a better price, but on the other side
he has not to pay penalties. This may work with an energy
exchange, but at the control power market the supplier is
pledged by contracts to be able to supply the scheduled power
at 100 % of the time he offers.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
87

III. SIMULATION SYSTEM

The basic simulation system consists of a wind farm with
250 MW and a storage system. These components were
modeled in a simulation environment, which was designed for
optimal operation planning of energy supply systems using a
mixed integer linear programming (MILP) [2]. The required
technical parameters of the wind energy farm and the storage
system are implemented into the model, whereby it is secured
that the facilities will run within their specifications. In this
work the program uses the optimization to make decisions
where the generated wind energy will be sold. For this purpose
different payment sources exist, which each represents another
customer (e.g. an energy exchange). For each scenario and
their combination the optimization tries to find the best result
by maximize the amount of coverage for the wind farm
operator. Equation (1) shows the objective function.

a minimum of 0.1 MW for 1 hour. The bidding is for the next

day and following the daily load, the price between noon and
afternoon is much higher than during night, so there can be a
worthwhile energy shifting.

B.
Control Power
To ensure a secure operation of the electrical grid, it is
necessary to use control power for the regulation of energy and
frequency fluctuations. There exist two types of control power:
first is positive control power which demands extra energy feed
into the grid (e.g. increased supply from power plants) and the
second is negative control power which reduces the energy
feed into the grid or increases the load (e.g. with pumped hydro
storages). The different types of control power (primary,
secondary, tertiary control power) are realized today mainly
with large conventional power plants.
These simulations contain only tertiary control power (TC).
After 15 minutes the TC has to be activated and takes place up
I T
to 1 hour. Suppliers have to deliver at least 15 MW and the
!
(1)
OF = R i,t Ci,t x i,t
MAX
bidding is for the next day. The transmission system operator
i =1t =1
purchases this power at a control power market, where
accredited power plants are able to offer their power. Wind
R
revenues
energy farms may also take part, because wind turbines are
C
costs
able to supply control power [5]. There exist already renewable
x
optimization variable
energy sources which deliver control power. Besides
i, I
index of number of elements
t, T
index of time intervals
hydropower, small and midsized power plants (wind,
photovoltaics, combined heat and power) working together in
The simulation expects a "day-ahead" sale of energy with generation pools to satisfy with the requirements.
The power plants get a fix amount of money only for
forecasts for wind & prices and makes a daily optimization
with 96 time intervals of 15 minutes each for one complete providing the power ready (demand charge). The price depends
year. The system uses also data from the next two days as input. on the demand at the power market; typically the negative
Thus, the usage of the storage unit can be more optimized for a control power is expensive during the night and cheap at the
longer time period. In the case the prices are higher the next day, while it is reverse at the positive control power. The
day, the optimization system maybe decides to store only special interest at negative control power for wind farms is the
energy to the storage at present day and sell it completely at the possibility to sell the reserved energy also at a source without a
contract, e.g. the EEG. As long as the negative control power is
next day. As input for the components historical time-variant
not required, the wind turbine can run on full power and sell
curves were used for price trends and wind speed. So it is
the energy. Only at the time of demand, the turbine must
guaranteed that the wind speed in the summer is corresponding
throttle the output and loss the extra income.
to the energy market price in the summer.

IV. BASIC SIMULATION DATA

For the simulations, scenarios were devised in order to
reproduce the energy market situation in Germany [3]. The
main focus in this work was on the energy and control power
market, but the model also included possibilities to participate
at other markets, e.g. congestion management [4].
A.
Energy Market
After the liberalization of the electrical grid the different
European energy markets arose. This scenario is based on data
from the European Energy Exchange (EEX). At the EEX
suppliers and customers can trade at different markets, whereas
in this work is considered only the spot market. There is the
possibility to offer hour-bids (in this work called "Peak") with

C.
Energy Storage
Energy storages are suited for making the energy supply
independent of time. They can accept and deliver energy from
the grid if there is a surplus or a shortage. Storages are not only
used for this technical issue but also for economical purposes.
During the night energy is charged into the storage and during
the high demand at the daytime it is feed back into the grid, e.g.
with pumped hydro storages. Thus, the expensive peak load
energy can be avoided partly.
For the simulations the storage is used both for technical and
economic issues. It charges the wind energy when the price at
the energy market is low and discharges when the price is high.
The storage also assures a secure supply of the scheduled
energy. According to the wind farm, a storage was chosen with
a power of 100 MW and 800 MWh (redox flow battery).

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
88

like before, because the energy is feed directly into the grid
without any contract like it is produced.
300

Power (MW)

D.
Further Specifications
The running costs (operating and maintenance) of the
storages were integrated into the optimization while fixed costs
for the storage or other expenses (e.g. fees for participation at
energy market) are not part of the optimization. They will be
offset later with the simulation outputs to get the overall result.
It is also assumed that the wind energy farm already exists, so
its costs are not part of this work.

200
100
0
-100

V. RESULTS

-200

300
200
Power (MW)

-300
0
nTCp
Wind

TnTC
nTCpH

P e a kS
BEP e a k

10

12

P e a kS h
EnTC B

14

16

18

EEG

20
22
Time (h)

24

Figure 3. At day of delivery differences occur between supply and sale

If there will be a safety factor, then the amount of maximal

possible power will be reduced. In Fig. 4 this value is reduced
by 15 %, which can be seen in the purple (PeakSF) and dark
green color (nTCSF). This energy can be sold also to the EEG
the next day or it will be later throttled if necessary. With this
"buffer" the wind farm owner can compensate differences
between the wind forecast from the day-ahead optimization and
the real wind. So on the day of delivery the missing energy
(Fig. 5) is reduced, compared to Fig. 3.
300

Power (MW)

The safety factor was varied between 0 to 20 % according to

the maximal possible power, as described in chapter II. With
these different values, the various scenarios were simulated for
one the same year to compare the results, according to
violation of the scheduled energy and also to the income of the
wind farm owner.
In Fig. 2, an example day of the day-ahead simulation is
displayed. The wind power (Wind) will be sold between 7.00
and 21.00 to the energy exchange, supplying always a constant
power block for one hour (Peak1, Peak2). In the morning, a
four hour block of negative tertiary control power (nTC1) will
be delivered, which will be provided by throttling the wind
farm (TnTC). The rest of the energy will be sold via the EEG
(EEG) with a payment of 80 /MWh. It is obvious, that the
optimization will sell as much energy as possible to the highest
price, so the maximal power is always the maximal wind
power, concerning to the wind forecast.

200
100
0

100
-100
0
-200
-100
-300
-200

0
nTC 1

-300
0
nTC 1

TnTC

4
P e a k1

6
P e a k2

8
EEG

10
Wind

12

14

16

18

20
22
Time (h)

4
nTCS F

P e a k1

10

P e a k2

12

14

P e a kS F

EEG

16

18

Wind

20
22
Time (h)

24

Figure 4. Optimized day-ahead sale of the wind energy

with a safety factor of 15 %

24

Figure 2. Optimized day-ahead sale of the wind energy

without a safety factor

300

Power (MW)

On the day of delivery, the wind speed differs from the

predicted wind the day before, so the produced wind power
does not fit the scheduled energy what was sold the simulation
before to the day-ahead markets. In Fig. 3, the sold scheduled
energy to the energy exchange (PeakS) and to the control
power (nTCp) cannot be delivered (PeakSh, nTCpH).
Therefore the missing energy has to be taken from other
sources (EnTCB). In these cases, the wind farm owner has to
pay a higher price to purchase this (balancing) energy (BEPeak)
or has to pay penalties to the costumer, because he could not
fulfill the contract. The rest of the energy is sold via the EEG

TnTC

200
100
0
-100
-200
-300
0

nTCp
Wind

2
TnTC
nTCpH

P e a kS
BEP e a k

10
P e a kS h
EnTC B

12

14

16

18

EEG

20
22
Time (h)

Figure 5. Supply with reduced violations due to the safety factor

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
89

24

Power (MW)

In the case a storage is combined with the wind farm, a

safety factor is not necessary because the storage will balance
the difference between forecast and real wind. Furthermore the
storage can increase the income of the owner because now it is
possible to speculate on the market as described before. The
same day as before is showed in Fig. 6. With the storage the
optimization decide to buy energy in the morning (EP) and
charge the storage (Sc). The storage is then discharged to
maximize the energy which is delivered to the energy exchange
(Sd) and it also supports the control power, which is increased
in the morning and evening by the storage (ScnTC).

TABLE I
OVERVIEW OF THE RESULTS FROM THE DIFFERENT SCENARIOS OF ONE YEAR

24 772 532

Results in Euro
Part of
Penalty
4 904 755

24 772 532

28 178 166

1 222 412

28 178 166

28 644 179

480 272

28 644 179

Violations
TC

Optimization

SF = 0 %

74x

SF = 5 %

27x

SF = 10 %

9x

Scenario

End Result

SF = 15 %

4x

28 582 357

254 864

28 582 357

300

SF = 20 %

1x

28 458 636

109 144

28 458 636

200

Storage &
SF = 0 %

37 330 143

3 359 631

27 330 143

100

VI. CONCLUSION

0
-100
-200
-300
0
nTC1
Sl

nTC2
EP

TnTC
Sd

10
ScnTC
EEG

12

14

P eak1
Wind

16

18

20

22

24

Tim e (h)

Peak2

Figure 6. Day-ahead optimization including the storage

At the day of delivery (Fig. 7) the storage is able to manage

the different wind speeds and no violation of the control power
occurs, only balancing energy for energy exchange is needed.
300
Power (MW)

balancing energy of the result and the last column shows the
total income including storage costs if included.

200
100
0

This work showed, that it is essential to use a safety factor in

the case, that the wind farm operator wants to participate at an
energy market without a storage. Under some circumstances it
is possible to allow violations of the scheduled energy, e.g. at
the energy exchange. Nevertheless, the suppliers have to
deliver the sold energy and so they have to buy it somewhere
else (for higher prices) or have to pay penalties. At the control
power market a non-compliance is not accepted, so in the case
the wind farm is not able to supply the sold energy, it is locked
out from further trade and cannot take part anymore at this
market. Therefore it is necessary to include a safety factor
(e.g. 20 %) to be on a safe side. However, including this factor
will decrease the income of the wind farm because of selling
less energy to the high price payment. An energy storage can
handle this problem and is also able to increase the income by a
better trading of the energy at the market. Thus, a storage has
both technical and economical benefits for the system [6].

-100

REFERENCES

-200

[1]

-300
0

10

12

14

nTCp

TnTC

ScnTC

P eakS

PeakSh

Sl

Sd

EEG

Wind

EPsh

BEPeak

B EEP

16
EP s

18

20

22

24

Tim e (h)

Figure 7. With a storage no violations occur and the income is increased

These simulations were made for each day of a chosen year

for the different scenarios. At the end, the income/expenses of
the wind farm owner including penalties and the annual storage
capital costs (10 m) have to be compared to give an answer to
the question, which safety factor (SF) is necessary and whether
the implementation of a storage is worthwhile. The results are
shown in Table I. The first column shows the number of
violations of the provided control power, because if there are
violations, the wind farm is banished from further
participations at this market. The next column is the results of
the optimization for one year; the part of the penalty fees and

[2]

[3]
[4]

[5]

[6]

"Act Revising the Legislation on Renewable Energy Sources in the

Electricity Sector and Amending Related Provisions (Renewable Energy
Sources Act)", Federal Ministry for the Environment, Nature
Conservation and Nuclear Safety, in: Federal Law Gazette 2008 I No. 49,
Bonn, 31 October 2008
P. Hacklnder, J.F. Verstege, "Accelerated MILP-Strategies for the
Optimal Operation Planning of Energy Supply Systems", in Operations
Research Proceedings 2002, Springer-Verlag, Berlin, Heidelberg (2003)
pp. 259-264
S. Vller, J.F. Verstege, "Investigation of new Payment Models for Wind
Farms in Combination with Energy Storages", in Proceedings of
International Youth Conference on Energetics, Budapest, Hungary, 2007
A.F. Kaptue Kamga, S. Vller, J.F. Verstege, "Congestion Management
with Energy Storages", accepted paper for CIGR/IEEE PES Symposia
Integration of wide-scale Renewable Resources into the Power Delivery
System, Calgary, Canada, 2009
A.-R. Al-Awaad, S. Vller, J.F. Verstege, "Contribution of Wind Power
Plants to Frequency Control", accepted paper for 4th International Exergy,
Energy and Environmental Symposium (IEEES-4), Sharjah, United
Arabic Emirates, 2009
S. Vller, A.-R. Al-Awaad, J.F. Verstege, "Benefits of Energy Storages
for Wind Power Trading", IEEE International Conference on Sustainable
Energy Technologies (ICSET), Singapore, Republic of Singapore, 2008

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
90

The Probabilistic Integration of Demand-Side Load

and Generation in a Representative Irish Distribution
Network
Keith Sunderland1
Dublin Institute of Technology, Ireland
[Link]@[Link]

Abstract-This paper introduces a methodology for investigating

the behaviour of a distribution network incorporating both load
variations and variable micro-wind generation output. The
probabilistic methodology is applied to a representative model of
the Irish Distribution Network. Application of a selection of
commercially available micro-wind turbines to the model is
investigated for a range of load and generation scenarios based
on a standard load profile and varying mean wind speed.
Analysis of the results of the analysis forms the basis for defining
the limitations of the methodology but also presents the
opportunities for future work.

I.

INTRODUCTION

This paper advances the work in [1], where a model was

developed to analyse the impact on a distribution network, in
the terms of voltage rise of increasing penetration of micro
generation and more specifically, micro-wind generation.
Indeed, [1] was itself based on the systems described in [2]
and [3]. In a report commissioned by Sustainable Energy
Ireland (SEI) investigating the costs and benefits of
embedded generation in Ireland [3], analysis for the context
of micro-generation stated that voltage rise problems would
not be encountered for the worst case scenario (each
consumer having 1.1kV of generation whilst at minimum
load of 0.16kVA).
This model is compiled employing MATLAB to
implement a Distflow routine as outlined in [4]. The Distflow
approach derives the bus voltage of a weakly meshed network
by iteratively evaluating the Load Flow solution of the
modeled network under varying load/generation scenarios.
The Representative Irish Network which is considered in
this analysis is shown in Figure 1. The system comprised of:

One 500MVA Source

Two 10 MVA Transformers (YY0, 38/10.5kV), autotapping to -20/+10% in steps of 1.67% and an AVC
scheme with bandwidth of 2.5%

Five, 10kV Distribution feeders with one modeled in
detail
o This feeder is 3km long with 1.5km being
185mm2 10kV PICAS and the remaining 1.5km
being of 95 mm2, 10kV PICAS

The detailed feeder contains ten 10/0.433kV
Substations (fixed tap transformer), each substation
having four LV feeders.

One of the LV Feeders is modeled in detail

Michael F. Conlon1
Dublin Institute of Technology, Ireland
[Link]@[Link]
o This feeder is 300m long with 150m being
185mm2 415V CNE and the remaining 150 being
of 95mm2, CNE.
The model developed in [1] was compiled for a range of
scenarios:

Maximum Demand on each substation (of each 10kV
feeder), 0% Generation

Minimum Demand on each substation (of each 10kV
feeder), 0% Generation

Minimum Demand on each substation (of each 10kV
feeder), 100% Generation
Each substation, serves 312 customers (over three phases)
and each customer is modeled as having an After Diversity
Maximum Demand (ADMD) of 1.28kVA and the potential of
affording a micro-generation technology with capacity of
1.1kVA. Each substation was modeled with a load factor of
50%.

500MVA Source

38/10kV 10M VA
Transform er (Auto-T ap)
(18% Reactance)

185m m 2 10kV PICAS

(length, 1.5km)
(0.164+ j0.08) O/km

Fifth 10kV Line

m odelled in detail

Five equal 3 km , 10kV

Feeders (Lum ped P aram eters)
com prising 1.5 km of
185m m 2, 3-core PICAS plus
1.5 km of 95m m 2, 3-core
PIC AS
Ten 400 kVA Substations
serving 312 consum ers
with ADM D of 1.28kVA
per consum er

95m m2 10kV PI CAS

(length, 1.5km)
(0.32+ j0.087) O/km
Four LV Feeders giving total
ADMD of 400kVA
300m 415V Fede r Ca ble:
1st 150m : CN E, (0.164 +j0.074) O /km (phase ); (0.164+j0.0 14) O /km (neutral)
2nd 150m : CN E, (0.32 +j0.075) O /km (pha se ); (0.32+j0 .016) O /km (neutral)

30m 35V Service C able (pe r

c onsume r):
CN E, (0 .851+j0.041) O /km (pha se );
(0.90+j0.041) O /km (neutral)

78 Consum ers equally

spaced over four sections

Figure 1: 38/10/0.4 kV Distribution Network Model

These scenarios represent the extremes of possibility. This

paper further develops this model by integrating variable load
and generation. This is implemented by using a probabilistic
approach to the modeling of the wind speed of the microwind turbines. This variable micro-wind generation is
combined with load profile data as provided in [6] which

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
91

represents standard domestic load profiles developed for the

Irish Electricity Market. These were utilized to determine the
variation in voltage profile for the LV network as a result of
both load and micro-wind generation. Similar work has been
undertaken for a Small Scale Energy Zone (representing a
total of 96 LV consumers) in [5] where a dynamic model
which integrated variable wind speed in terms of a minimum
load of 0.16kVA and a percentage of customers having a
1.1kW generator connected, produced results suggesting that
voltage tolerance is breached at 114% penetration level.
The approach taken here is to apply a single (but variable)
wind speed to all connected generation. This is considered
acceptable as it is assumed that there would be little variation
of wind in the area (excluding the effects of obstruction or
shading) encompassed by the representative Distribution
Network. Independently varying bulk connected load of the
connected consumers at the sectionalized connections of the
LV feeder was also considered acceptable as a compromise to
individual manipulation of each consumers load.
Analysis of voltage profile over the course of
representative days in different seasons was implemented for
a sample of commercially available micro-wind generator
units relative to the load profile.
Distflow is employed as the optimal methodology to derive
the load flow results for the network as standard load flow
approaches (Newton Raphson, Gauss Seidel etc.), have
difficulties in converging due to the high X/R ratios
characteristic of distribution networks.
II.

V (i ) = VMean

log (1 F )

(1)

where:

V(i) is speed derived for each iteration of the program

and N is the total number of time periods considered.
The analysis is carried out with a 15 minute interval to
align with the load model data. F is a uniformly
distributed random number. The other assumptions
made in relation to the analysis are as follows:
Uniform penetration of a chosen wind turbine across
the network
o Generation is lumped to the respective
buses, but contrary to [1], this lumped
generation is variable with respect to the
applied simulated wind speed.
The chosen wind turbine is modeled in terms of
o Cut-in/out speeds
o Rated Power (at rated wind speed)
o Actual power output is based on the derived
Rayleigh distribution in conjunction with
the specific mean speed
The load data acquired was for a representative year
based on 15 minute intervals [6].
o Figure (2) illustrates the load variation over
the year.
Automatic tapping of the 10 MVA Transformers is
not implemented.

THE METHODOLOGY

Figure 2 illustrates a flow chart describing the methodology

utilized in deriving the bus voltages for the network model

Figure 2 shows the load variation over the course of one year,
with a peak demand of 1.73kW and an annual consumption of
6000kWh.

R andom
N u m be r

2
1.8

R A Y L E IG H
D IS T R IB U T IO N

1.6

W IN D P O W E R
(E le c tr ic a l)

L o o p in
te r m s o f
L o a d P r o fi le

L o a d P r o file
(P e r io d /Y e ar )
NETW ORK
PARAM ETERS

D IS T F L O W
(L o a d flo w )

Demand, kWh for 15 mins

B u s /B r a n c h
In fo r m a t io n
(In itia lis a ti on )

1.4
1.2
1
0.8
0.6

N o?

C o n v e rg e n c e ?
Y es?

BUS VOLTAGES

0.4
0.2
0

50

100

150

200

250

300

350

Day
NETW ORK
VOLTAGE
RECORD

Figure 3: Load Profile Representation of Irish Domestic Dwelling

Figure 2: Flow Chart illustrating the methodology in deriving a load flow
solution for variable load/gen

The approach is summarized as follows:

A random number is generated in MATLAB and
through (1) a Rayleigh probability distribution is
created to represent a varying wind speed.

III.

ANALYSIS

Three different types of commercially available wind

turbines (Table 1) were modeled for the following scenarios:
o Load/Gen at unity power factor (p.f.)
o Load @ 0.95 p.f.; Gen @ unity
o Load@ unity; Gen @ 0.95 p.f.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
92

o Load/Gen @ 0.95 p.f.

Two periods over the representative year were
considered in detail:
o The 5th of January and 5th of July, with
mean wind speeds of 5.49m.s-1 and
3.65m.s-1 respectively [7]
o The analysis is performed in terms of the
bus at the end of one of the LV feeders

5th January: 100% Generation (unity p.f.)

3

14

Lo ad (kW)
Gen 2 (kW)

2.5

Generation & Load (kW)

TABLE I

12

Gen 3 (kW)
Gen 4 (kW)

10

W ind Speed
8

1.5
6
1
4

0.5

SAMPLE OF COMMERCIALLY AVAILABLE WIND TURBINES

0

Model
Whisper
200
Proven
2.5
Swift

Reference

Cut-in
Speed

Cut-out
Speed

Rated
Speed

Rated
Power

Gen 2

3.58 ms-1

>50 ms-1

14 ms-1

1.5 kW

Gen 3

2.5 ms-1

>50 ms-1

12 ms-1

2.5 kW

-1

14 ms-1

1.5 kW

-1

Gen 4

3.58 ms

>50 ms

0
00

04

08

12

16

20

Figure 4: The relationship between generation and wind speed reflected in

terms of the Demand profile
5th January: 100% Generation (unity p.f.)
[5th Jan] Gen 2:Load/Gen = Unity

1.15

14

[5th Jan] Gen 3:Load/Gen = Unity

[5th Jan] Gen 4:Load/Gen = Unity

IV.

SIMULATION RESULTS

1.1

12

Wind Speed

Voltage (pu)

10

The simulations performed, as summarised in Table 2,

were in the context of voltage profile across the entire
network, but for simplicity, the voltage at the remotest bus on
an arbitrary LV branch (400V) was recorded. The voltage
magnitude for this bus is measured against the tolerance of
+10/-6% of nominal voltage for the Distribution Network.
The 5th of January and 5th of July were specifically chosen on
the basis that there would be considerable differences to load
and generation profiles due to both the associated mean wind
speed as well as load profile.

1.05

4
0.95
2

0.9

0
00

04

08

12

16

Figure 5: Voltage variation with respect to wind speed for the 5th January
5th January: Voltage Profile Summary
1.3

TABLE 2

[5th Jan] Gen 2:Load /Gen = Unity

1.25

VARYIABLE LOAD AND GENERATION SIMULTION SUMMARY

[5th Jan] Gen 2: Gen= 0.95; Load =

Unity
[5th Jan] Gen 3:Load /Gen = Unity

1.2

th

th

5 January & 5 July

Generation

100%

Unity

Unity

Generation

0.95

Unity

Connection

Unity

0.95

[5th Jan] Gen 3: Load = 0.95; Gen =

Unity

1.15

Voltage (pu)

Load

20

1.1
1.05
1
0.95

50%

Unity

Unity

0.9

Generation

0.95

Unity

0.85

Connection

Unity

0.95

00

04

08

12

16

20

Figure 6: Voltage Profile for all cases during the 5th January

The results of the simulations are illustrated in the

following graphs. Figure 4 illustrates the relationship between
the simulated wind speed and the generation of 5th January
with 100% generation at unity pf. The typical daily load
demand as obtained from [6] is also shown. The different
levels of output from the modeled generators are clearly
discernable with the Proven 2.5 having the biggest
contribution. As the rated outputs of the Whisper and Swift
Models are closely aligned 1kW and 1.5kW respectively, the
generated outputs from both models are very similar at given
the wind speed profile. Fig. 5 shows the voltage over the
course of the day which tracks the simulated wind speed.

Fig. 6 illustrates the voltage profile for the 5th of January.

The same analysis was performed for the 5th of July
producing similar results albeit with a lower voltage
magnitude due to the lower generation for the date. There are
increased levels of generation on the 5th of January due to the
increased mean wind speed being employed and as a result,
there is a higher voltage level experienced by the bus under
scrutiny for the demand applied.
Figures 7 and 8 depict the voltage spread for the varying
load/demand ratios for the respective dates under scrutiny.
The minimal impact afforded by the generation due to the
lower mean wind speed is particularly evident in Figure 8.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
93

Also, the lower consumer demand is also evident in that the

breaches of the lower voltage threshold are less frequent and
their level diminished.
Frequency

PDF of Voltage Profile wrt to Generator Choice

Voltage Spread : 5th January [Varying Generation]

Pr oven 2.5

50%

Swif t

Whisper 200

40%

Wind Sped

Mean Wind
Speed = 6 ms-1

30%
20%
10%

[ 50%Gen] UNDER %
50

60%

0%

[ 50%Gen] OVER %

0.94

[ 100%Gen] UNDER %

0.96

0.98

1.02

1.04

1.06

1.08

1.1

1.12

1.14

Mor

V o l t ag e ( p u)

[ 100%Gen] OVER %

30

Figure 9: PDF of Voltage Proliferation in terms of the different Generators

% Deviation

10

CONCLUSION
G en 2 :
Lo ad / G en
= U ni t y

-10

-30

G en 2 :
Gen=
uni t y;
Lo ad =
0 .9 5

G en 2 :
G en=
0 .9 5;
Lo ad =
U ni t y

G en 3 :
Lo ad / G en
= U ni t y

G en 3 :
G en=
uni t y;
Lo ad =
0 .9 5

G en 3 :
G en=
0 .9 5;
Lo ad =
U ni t y

-50

-70

Figure 7: Voltage Spread for all contexts during the 5th January

40

Voltage Spread : 5th July [Varying Generation]

[ 50%Gen] UNDER %
30
[ 50%Gen] OVER %

[ 100%Gen] UNDER %
[ 100%Gen] OVER %

% Deviation

20

10

The work presented introduces a methodology that can

investigate distribution network behavior when incorporating
both load and generation variation. A sample of commercially
available micro-wind turbines were utilized in a
representative distribution network model with their outputs
based on their characteristics and wind speed derived from a
specified probability distribution. This model was subjected
to mixed generation and loading conditions and demand was
based on a standard load profile for domestic consumers.
Voltage variation was monitored. Voltage is dependent on
both load as well as generation, with voltage rise resulting
from reverse power flows from the connected generation.
Generation based on wind will result in fluctuations
mathching the turbulent nature of the wind speed. The model
had no voltage control on the variable tap distribution
transformers and the simulations illustrated this by virtue of
both the frequency of voltage tolerance breach and the spread
of voltage for the different load/gen scenarios.
Indeed, voltage control by virtue of transformer tapping
presents a considerable challenge for demand side
management due to the stochastic and variable nature of
wind.

Gen 2 :
Lo ad / Gen
= U nit y
- 10

Ge n 2 :
G en=
unit y;
Lo ad =
0 .9 5

G en 2 :
Gen=
0 .9 5 ;
Lo ad =
U nit y

Gen 3 :
Lo ad / Ge n
= U nit y

G en 3 :
G en=
unit y ;
Lo ad =
0 .9 5

Gen 3 :
Gen=
0 .9 5;
Lo a d =
U ni t y

Figure 8: Voltage Proliferation for all contexts during the 5th July

Figure 9 shows the impact on the voltage level with respect to

the range of generators. This is placed against the background
depicting a probability density function describing the wind
speed for the year with mean 6ms-1. Clearly, the Proven 2.5
with its greater output contributes most frequently to
breaching the upper voltage limit of 1.1pu. The graph also
illustrates that for each of the wind turbines, the most
frequent voltage is 1.08pu, which could be considered
excessive and this is irrespective of the wind turbine chosen.
The mean wind speed of 6 ms-1 was chosen as a more
realistic year-round average.

REFERENCES
[1]

[2]
[3]
[4]
[5]
[6]
[7]

Sunderland K., Conlon M.F. The Role of Microgeneration in Irelands

Energy Future, Universities Power Engineering Conference, 2008,
UPEC 2008
Ingram. S., P.S.a.J.K., Impact of Small Scale Embedded Generation on
the Operating Parameters of Distribution Networks, Department of
Trade and Industry. June 2003.
SEI, Cost and Benefits of Embedded Generation in Ireland, Report by
PB Power, 2004
S. Ghosh, D.D., Method for load-flow solution of radial distribution
networks
Cipcigan L.M, Taylor P.C and Trichakis P. The Impact of Small Scale
Wind Generators on LV Distribution System Voltage
[Link]
MET Eireann, 30 Year Climatic Averages,
[Link]

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
94

Planning of Ancillary Services Securing Power

System
Ondej Novk1), Tom Strnad1), Petr Horek1), Josef Fantk2)
{novako5,strnat2,horacek}@[Link], fantik@[Link]
1)

Czech Technical University in Prague, Faculty of Electrical Engineering

2)
EPS, a.s., Czech Transmission System Operator

Abstract- Procuring sufficient amount of operational reserves

used by the Transmission System Operator (TSO) for keeping a
balance between generation and load in real-time is a complex
task. Extensive balancing power reserves enable keeping the Area
Control Error (ACE) close to zero most of the time but the cost of
enabling these reserves might be very high. The paper proposes a
method of planning sufficient amount of operational reserves or
rather Ancillary Services on economical basis. The goal of the
presented algorithm is to find AS combination that minimize the
cost of AS reservation and at the same time secures transmission
system operations. Monte-Carlo simulation scheme of the area
operation provides sufficient insight to ACE behavior under
different operational reserves. Branch and Bound search scheme
is applied to arrive at the best compromise between power
security and cost of AS procurement. Short term, one month
planning horizon is considered.*

I.

INTRODUCTION

Prior to the restructuring and liberalization of the European

electricity industry responsibility for security of supply was
generally the role of vertically integrated companies with an
obligation to supply. Unbundling of the activities has resulted
in the establishment of Transmission System Operators (TSO)
and changes in roles and responsibilities.
TSO companies from 34 European countries created a new
association, the European Network of Transmission System
Operators for Electricity (ENTSO-E) in December 2008. The
ENTSO-E structure includes three committees for panEuropean activities on System Development, System Operation
and Market Frameworks being also the main responsibilities of
a TSO. TSO is the ultimate instance in the chain of electricity
market players participating to active power balancing.
Electrical energy market players are economically responsible
for meeting their planned individual balances in each
settlement time frame. While market players are commercially
responsible for their actions in all timescales, after gate closure
the TSOs are physically responsible for securing power system
operation and maintaining balance between supply and demand
by having access to sufficient operational reserves ( Fig.1).
Balancing power is a tool used by TSOs to maintain the
instantaneous physical balance. Balancing power is a
*

This work was supported by the Ministry of Education of the Czech Republic
under the project 1M0567 and by the Czech TSO under the project Reliability
and economy of system services

commodity which is made available on the market in the form

of Ancillary Services (AS). One of the tasks of the TSO is to
plan and purchase enough AS and activate them timely to meet
the grid performance standards and minimize the cost of the
AS purchase taking into consideration all physical and market
constraints. A TSO might define and use different AS
depending on the control area/country specificity of generation
sets, however, there are some generic categories present in
every control area: Continuous Regulation, Energy Imbalance
Management, Replacement Reserves, Voltage Control and
Black Start. The paper deals with procurement of all AS
categories except voltage control and black start as these do not
contribute to power balancing directly.
Performance standards are specified and set for a control
area of the UCTE following general requests for all
interconnected areas defined in the UCTEs Operation
Handbook [1]. An important aspect, which influences the
choice and volume of AS, is the cost of the purchase which
should be kept low. Simulated operation of the control area
under the TSOs Automatic Generation Control and Energy
Management System dispatching balancing power in real time,
as shown in Fig.1, will be used for generating time series of
To plan and purchase
factual
frequency bias obligation
inadvertent interchange

scheduled export

balancing power
from other areas

hours

f
K

[Hz]

net scheduled interchange

Pprogrammed
scheduled import

MW

f0

range of balancing power reserves

(purchased ancillary services)
fmeasured
f0

power reserve activation primary frequency

primary frequency
control
control

TSO/ISO
_
+
_

ACE
[MW]

PIcontroller
controller
AGC
PI

auto

man

set-point of foreign exchange

(scheduled exchange)

balancing
balancing
power
power

area
area
generation
generation

[MW]

EMS

+
production

_
load shedding
shedding
load

BA's ACE
+
-

unscheduled import
unscheduled export

net actual interchange

metered

area demand
demand
area

[MW]

consumption

Pactual

Figure 1 Power balancing in the area as a feedback control system

Area Control Error (ACE) for one year areas operation.

Every TSO might adopt different AS planning procedures
depending on interconnection and control area specificity.
What make the problem more difficult are recent changes in
market structures so there is a lack of long term experience in
effective AS planning. The principles presented in the paper

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
95

might serve as a basis for a general platform adopting various

criteria and constraints despite the fact that the approach was
developed for one control area of the UCTE interconnection.
In chapter II problem definition is introduced. Chapter III
describes proposed solution. In chapter IV achieved results are
shown. Final chapter deals with conclusions and outlook for
future work.
II.

PROBLEM STATEMENT

Goal of presented work is to find AS combination with the

least reservation cost that secures reliable transmission system
operation. From optimization point of view the problem can be
stated as a reservation cost minimization constrained by
reliability standards that have to be met.
Reliability standards of the AS combination are evaluated by
means of Monte Carlo simulation which incorporates
transmission system statistical model parameterized by values
obtained from processing of real operational data of the Czech
transmission system and dynamic model of the AS activation.
The list of AS considered is given in Tab. I as defined by the
Grid Code of the Czech TSO [5]. Details of how the Simulator
is designed and used in the Monte-Carlo scheme to calculate
time series of ACE can be found in [4].
A. Ancillary services definition
In the planning algorithm five types of AS are considered:
+
RZSR, RZTR , RZTR , RZQS and RZDZ. Note that the method
allows working with any other AS. AS definition changes from
time to time and the algorithm will not be practical is restricted
to particular AS only.
TABLE I

B. Reliability standards
Result of Monte-Carlo simulation and control performance
of used AS combination is expressed by area control error
which is given:
ACE = P + K f ,

where P is an inadvertent interchange (measured difference

between actual and scheduled interchange with neighboring
areas), f is a frequency error (difference between measured
frequency factual and the frequency set-point f0) and K is a
frequency bias of the control area. Formula (1) matches Fig.1
except for sign (polarity) of ACE. Fig.1 respects the way how
P is calculated by Czech TSO. By statistical processing of
ACE reliability indices are obtained. Definitions and
descriptions of reliability indices used by Czech TSO are
shown in Tab. II.
TABLE II
RELIABILITY INDICES
Index
name

Index
unit

rACE1

rACE60

nACE1t

E2100

MWh2

ACE

MW

ANCILLARY SERVICES

ACE

MW

Response
time

ACE

MW

E100

MWh

Ancillary Service
RZPR (spinning)

Description

Guaranteed

primary control

Yes

~ 30 sec.

RZQS (non-spinning)

quick-start reserve
(pumped-storage)

Yes

max. 10 min.

RZSR (spinning)

secondary control

Yes

max. 10 min.

RZN30+ (nonspinning)

RZTR+ (spinning)

stand-by reserve

Yes

load change
emergency
assistance from
abroad
tertiary control

Yes

60

load change

Yes

RZN30- (nonspinning)

emergency
assistance from
abroad

No

RZTR- (spinning)

tertiary control

Yes

RZDZ (nonspinning)

stand-by reserve

Yes

EregZ (non-spinning)

balancing energy

No

max. 30 min.

Description
Probability that the absolute average value of
one-minute (clock minute) averages of ACE
exceeds 100 MW over given period.
Probability that the absolute average value of
one-hour averages of ACE exceeds 20 MW over
given period.
Number of events when ACE remains
higher/lower +/- 100MW for time longer than 15
minutes in given period.
Sum of squared energy of events when ACE
remains higher/lower than +/- 100MW for longer
than 10 minutes in given period.
Average value of one-minute averages of ACE
over given period.
Standard deviation of one minute averages of
ACE over given period.
Standard deviation of one hour averages of ACE
over given period.
Sum of absolute values of energy of events when
ACE remain higher/lower than +/- 100MW for
time longer than 10 minutes in given period.

C. Reservation cost evaluation

AS reservations cost evaluation is modeled by elasticity
curves which respect relation between price and AS bids so
that mean unit price for reservation is increasing with
increasing amount of AS. Total reservation cost is given by

No
Yes

(1)

TC =

RZxx

RZxx

( RZxx ) RZxx ,

(2)

where RZxx stands for ancillary services RZSR, RZTR+,

RZTR-, RZQS and RZDZ respectively and CRZxx denotes
unit mean reservation cost of RZxx given by its elasticity curve.
more than 30
min.

D. Reliability indices computation

Reliability indices for selected AS combination are generated
by means of Monte-Carlo simulation [4]. Two hundreds of
simulation runs were simulated for each AS combination to
arrive at statistically better conditioned results. The core of the
proposed method is the way how AS combination is selected.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
96

III. BRANCH AND BOUND ALGORITHM DESCRIPTION

In this chapter a branch and bound algorithm used in the
procedure of generating and selecting prospective combination
of ACE is be described. Principal scheme of the proposed
algorithm is shown in (Fig. 2).

Figure 2 Principal scheme of the algorithm

Before starting the search algorithm several initialisation

steps have to be completed. The search space of all AS
combinations to examine is defined first. As only discrete
values of reservation are expected, all AS reservations are
separated into distinct values. Other constraints on reservation
may be applied, e.g. minimum recommended volume of
secondary regulation specified in [1] or maximum volume of
AS certifications in regulation area. Step size and limits used in
our studies are show in Tab. III.
TABLE III

Reliability indices
Ancillary
service
RZSR
RZTR+
RZTRRZQS
RZDZ

Lower bound
[MW]
200
0
0
300
0

Upper bound
[MW]
700
700
400
1000
200

Step size
[MW]
50
50
50
100
50

The next step is to specify reference value of reliability

indices as a minimum requirement we would like to achieve.
Reliability indices can be given directly or by means of
reference AS combination. In that case reliability indices are
determined by Monte-Carlo simulation with the reference AS
combination.
Final step of initialization is to choose initial AS
combination. If the reference combination is used then it is also
taken as initial AS combination otherwise random selection
applies.
A. Reliability indices evaluation
Next step after initialisation is generating ACE in MonteCarlo simulation scheme of the control area operation and

getting the expected values of reliability indices from statistical

processing of ACE time series.
B. Search space pruning procedure
Search space pruning depends on whether or not reliability
standards are satisfied. In the first case we obtain upper cost
estimation of optimal solution and hence every AS
combination with reservation cost equivalent or greater than
current upper bound can be removed from search space.
In latter case pruning algorithm is based on presumption of
lower amount AS reservation, that is needed to assure reliable
control area [Link] if the actual AS combination does
not satisfy reliability standards, it is natural to expect that any
combination with less reserves will not satisfy reliability
standarts either, and can be deleted from the search space. To
show that previous idea is correct the concept of monotonicity
needs to be introduced. A reliability index is monotone in some
AS if lower amount of the AS reservation implies higher value
of the index and consequently less reliable control area
operation. If monotonicity holds for all reliability indices the
pruning method described above can be used. Due to the
complexity of the modeled control area and the indices
definition, the relation between the amount of AS and the value
of indices is difficult to express analytically. Therefore
simulation approach was employed to find indexservice
combinations where monotonicity does not hold:
rACE1 index does not preserve monotonicity in RZTR+
and RZTR-,

(3)

rACE60 index does not preserve monotonicity in RZSR,

(4)

E2100 and E100 do not preserve monotonicity in any AS,

(5)

for the rest of reliability indices monotonicity holds.

(6)
Therefore any AS combination for which monotonicity holds,
and volume of AS is lower then the actual AS combination, can
be remove from the search space.
C. Selection of new AS combination to examine
The cheapest AS combination found so far is returned as a
solution when the search space is empty the search algorithm
stops, otherwise we are looking for the new AS combination
which will be examined in the next iteration. As in the previous
step of algorithm, the AS selection depends on whether or not
reliability standards are satisfied. First of all the next AS
combination with lower amount of reserves in less utilized
service is selected. If such combination does not exist in search
space, AS combination with lower amount of the most
expensive service is selected. If such combination does not
exist the next AS combination is selected randomly. In the case
when the reliability indices are not satisfied AS combination
with higher amount of the most utilized service is selected. If
such combination does not exist the next AS combination to
examine is selected randomly.
By selecting the next AS combination the algorithm
continues in next iteration as described in the paragraph A.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
97

A. Case study #1
The control area simulator is fed with three average monthly time
series of ACEO, where ACEO is the expected area control error
as would be without TSOs regulation reserve activations.
Selected ACEO are average in sence of indices values
calculated from ACE after TSO's regulation using the reference
AC [Link] inputs were chosen from 200 randomly
generated ACEO that respect statistical behavior of the Czech
control area.
B. Case study #2
Seven monthly ACEO time series equidistantly chosen from
the same 200 time series as in the case #1, ordered by
reliability indices computed with reference AS combination,
were used.
Resulting reserves and the relative cost of their procurement
in relation to the reference is shown in Tab. IV and Tab. V
respectively. Reliability indices of the solutions are given in
Tab. VI and Fig. 3.
TABLE IV
FOUND AS COMBINATIONS
Ancillary services combination

Reference
Case #1
Case #2

RZSR
[MW]
290
340
280

RZTR+
[MW]
260
200
100

RZTR[MW]
100
100
150

RZSQ
[MW]
570
600
700

TABLE VI
RELIABILITY INDICES OF THE FOUND AS COMBINATIONS
Reference
rACE1
rACE60
Reliability indices

In this chapter simulation results are presented. The

algorithm was used in month-ahead AS planning. AS
combination from year-ahead plan was taken as a reference.
Reference reliability indices were computed by means of
Monte-Carlo simulation of the control area with two hundreds
of monthly runs using reference AS combination.
The algorithm performance is shown in two case studies.

2,61
2,49

2,68
2,80

0,07

0,05

0,06

[MWh2]

523 265

1 225 797

1 619 786

[MW]

-0,63

0,15

0,42

ACE

[MW]

45,39

46,17

46,88

ACE

[MW]

11,69

13,04

14,12

E100

[MWh]

25 715

25 766

25 930

350
Case #1
Case #2

300
250
200
150
100
50
0

RZDZ
[MW]
160
0
100

-50
-100

rACE1 rACE60 nACE1t E2

100

ACE

ACE

ACE

60

E100

Figure 3 Relative values of reliability indices

REFERENCES
[1]

Reference

Case #1

Case #2

100,00

99,59

97,75

[2]
[3]

[4]

V.

2,65
2,86

Case #2

2
E100

60

TABLE V
RELATIVE COST OF THE FOUND AS COMBINATIONS

Relative cost of reservation [%]

%
%

Case #1

nACE1t

ACE

Reliability indices value [%]

Reference = 100%

IV. SIMULATIONS RESULTS

difference in accuracy between indices evaluation during

planning procedure and in evaluation of the found solution by
means of two hunders of Monte-Carlo simulation runs.
Minimalisation of this difference is main goal of the future
development of the method.

CONCLUSION

Developed algorithm was tested on month-ahead ancillary

services planning problem. Ability of presended method was
proven by two case studies show the ability of the algorithm of
finding better solution than was the reference. However, not all
indices were satisfied as can be seen (Fig. 3) that is because of

[5]

UCTE Operation Handbook, v2.2/20.07.04, Union for the Co-ordination

of Transmission of Electricity, Brussels (2004)
Reliability Standards for the Bulk Electric Systems of North America,
North American Electric Reliability Corporation, Princeton, NJ (2008)
P. Havel, P. Horek, V. ern, J. Fantk: Optimal Planning of Ancillary
Services for Reliable Power Balance Control, IEEE Trans. on Power
Systems, Vol. 23, No. 3, 1375-1382 (2008)
V. ern, , P. Janeek, A. Fialov, J. Fantk: Monte-Carlo simulation
of electricity transmission system operation, in Proc. 17th IFAC
World Congress, Seoul, Korea, 2008.
Rules for the Transmission System Operation Extract from the Grid Code,
Revision 09, EPS a.s., (January 2009) , [online],
[Link]

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
98

Biofuels- the power from plants

Grzegorz Maliga a, Amar Patil b
a

Phds student of Departament of Mechanical and Power Engineering, Wrocaw University of Technology
[Link]@[Link]
b
MScs student of Departament of Computer Science, Wrocaw University of Technology
amar.ec006@[Link]
Abstract

In this paper I try to review five fundamental

groups of biofuels which are used as a power source for
mechanical engines. However engines could be use as a
propulsion in cars or other machines and as a current
source when engines cooperate with generators.
The basics biofuels which belong to the fundamental
group are: rapeoil, biodiesel, biomethanol, methan from
biogas and synthetic biofuel. I present calorificvalue of
these kind of biofuels and compare to earthfuels.
In Europe the leader of production biofueals is
Germany. On the basic of this country I present how
the biofuel industry deweloped in the last ten years and
what kind of benefits is possible to gain. I describe also
political and law conditions which influece by using
biofuels.
Finally, I pay attention to the most important result of
using biofuels which is influence to environment,
because producing and burning biofuels take a part in
the same carbon dioxide cycle.
Keywords: biofuels, biodiesel, biomethanol, rapeoil,
fuels from biomass

I. INTRODUCTION
At present we have an awareness that the sources
of fossil energy are becoming lower. In order to stop
the disadvantageous phenomenon of decreasing raw
fuels and changing climate, we need to start using
gradually a new sources of renewable energy.
Renewable energy was common used till the 18th
century from that supplies as wind and wood. After
the industrialisation changes and impoves people
started used fossil sources of energy. In the 19th
century the renewable sources of energy were
comletely replaced by coal, oil and earth gas. These
changes were as a consequence of huge industrial
growing which reqired a more efficient and high
calorific value sources of energy than wood or wind
energy. Today result is that 90% of energy which we
use becomes from fossil resources. Using and
burning coal, oil and earth gas influences very
incovenient on climate and environment because
producing energy from raw fossil fuels releases great
size of carbon dioxide into atmosphere in short time,
which results in greenhouse effect. On possible
solution to decreas greenhouse effect is becoming
greater of using fuels, which producing and burning

take part in the same dioxide cycle. These kind of

fuels are producing from plants and generaly are
divided in two groups such as: biomass and biofuels.
Biomass characterises as raw kind of fuels for
example wood, straw or pellets, which are generally
burned in furnace in order to produce heat.
Whereas biofuels describe as kind of fuels which are
producing from raw biomass such as corn or plants
wastes in enrichment process. To biofuels group
belong: rapeoil, biodiesel, biomethanol, biomethanol
and synthetic biofuel. The bigest advantage of
biofuels, compare to biomass, is high level of
calorific value sometimes similar to fossil fuels.
Current life standarts require from people high
mobility, which influences on using some kinds
means of transport such as cars, trucks or others. The
common feature of propulsion machines is
converting one form energy (for example chemical)
to mechanical, which occur in internal combustion
engines. In order to acces satisfactory efficient and
working opportunities, are used fuels whch have
clorific value on level about 40 MJ/kg such as diesel,
petrol or liquid petrol gas. These kind of fuels can be
replaced only by biofuels, because the calorific value
level is comparable and they can be adapt by modern
engines developments.
Biofuels are also easy to storage and management,
which means that new technoloy, high investment
and infrastructure are not necessary for using
biofuels. For these reason in Germany in 2005 the
share of biofuels was around 3.6 % of total fuel
consumption. The large part of these biofuels is
biodiesel.
In the future for become greater the producing
of biofuels, larger areas have to be cultivated than
now. Interesting is also aspect how is the impact
increasing cultivated areas for energy on restrict food
production. Whereas the productivity of agriculture
has risen constantly and the trend will maintain in the
future. For example in Germany the share of freed
land for non food production has increased [1]. In
2005 over one million hectares were cultivated for
energy production from rapeseed, rye and wheat.
Specialist predicted that till 2030 the area for energy
production will approach to four million hectares.
The predictions also shows that in Germany the
biofuels share of total fuel supply in 2020 can
achieve 25%.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
99

II. BIODIESEL

III. BIOETHANOL

The most widespread of biofuels is biodiesel,

because it is adapted to the diesel engines. For
example in Germany biodiesel contributes around 18
million tons per year to the fuel consumption [1].
From 2004 the mineral oil corporation add biodiesel
to original diesel by 5 %, which allows to carry out
technical requirements for vehicle owners.

Biodiesel and vegetable oil are useful by diesel

engines but bioethanol can replace petrol. Bioethanol
can be mixed with petrol from mieral oil
manufactures in share up to 5 %. For use more share
of bioethanol the vehicles should be adapted.
Ethanol is produced by fermenting the sugar
included in plants. The sugar stach and cellulosebearing occur mainly in wheat, rye, maize and sugar
beet. In the future by developement of well-fitted
enzymatic processes also the wood, and straw could
be fermented.

TABLE I
BIODIESEL BASIC INFORMATIONS [1]
Raw materials:
Annual yield per hectare
Fuel equivalent
Market price
CO2 reduction
Technical information

Rapeseed or other vegetable

oils
1550 l/ha
1l biodiesel substitutes 0.9 l of
diesel
0.75-0.90 EUR/l
Approx. 70 %
Biodiesel in pure form:
manufacturers
approval
required; mixtures up to 5 %
without refitting the engine

From chemistry point of view biodiesel is a

vegetable oils methyl ester or fatty acid methyl.
Biodiesel production behaves by ester intercharge of
vegetable oil. For this reaction the methanol is
required and mixed in a ratio o 1:9. In temperature of
50 to 80C a catalyst is added in quantity around 0.5
to 1 %. The nature of reaction which occures , the
vegetable oil molecule composed of glicerine and
three fatty acid chains is broke down. The fatty acids
combine with the methanol to form biodiesel.
Glicerine is produced and later used in many fields
such as pharmaceutical or food industry.

Figure 1. Ester interchange of rapeseed oil to create biodiesel [1]

Using the biodiesel to the engine is depended on

viscosity and ignition properties, which are similar to
fossil diesel. By adding some components the ability
of using in winter is possible witout difficult to minus
20C. Important is also the lubrication effect which is
higher than fossil fuels. Consunption of biodiesel is 5
% more then diesel because the yiel energy is lower.
The requirements necessary for the fuel quality
are defined in the Europe standard DIN EN 14214.

TABLE II
BIOETHANOL BASIC INFORMATIONS [1]
Raw materials:
Annual yield per hectare
Fuel equivalent
Market price
CO2 reduction
Technical information

Grain, sugar
2560 l/ha
1l biodiesel substitutes 0.66 l of
petrol
0.50-0.60 EUR/l

30- 70 %
Can be mixed with fuel by up to
5%

The raw plants which included sugar are fermented

by yeast and enzymes to ethanol. When plants
containe the starch, that first the starch is converted
into sugar by enzymes. From fermentation proces the
by-product which occures in large quantities can be
used latter as fodder for the biogas plants.
Properties of ethanol improve the quality of
petrol, because the alcohol has a higher octane value
than petrol. When the octane value is higher , it
means the better antiknocking property. But the
disadvantage of bioethanol compare to pertrol is
calorific value, which is one third lower than
conventional petrol. After mixing ethanol increase
the vapour pressure of fuel.
Since 2005 in Germany many series vehicles are
available to run with ethanol proportion up to 85 %,
which is called flexible-fuels.
Mixing ethanol with petrol is restricted by petrol fuel
standard DIN EN 228 permits mixing of up to 5 %
of ethanol with petrol. For example when the 5 % by
volume of all petrol fuels were to be replaced by
ethanol, this would be around 1.3 million tons per
year.
IV BIOMETHANE
Biomethane is called as a biogas, which is for
example gained in agricultural facilities, in general
by the fermentation of maize silage or manure. To
produce biogas can be used many kinds of organic
substrate. Farms, which basen on cattle or pigs
husbandry, can use animal excrement, as a main
material for anaerobic fermentation. On the othre
hand there are a lot of farms where only plants are
used for energy production. In this kind of farms they
exploid mew renewable materials as crops, grass,

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
100

maize, sunflawers and others. To increas biogas

production and to utilize non-agricultural but organic
wastes some biogas factories use the residues from
food industry such as vegetable from wholesales,
distillers wasch or grease. The process of residues
fermantation provide many benefits, because the
dealing cycle of waste food is closed and it is
hygienic.
In Germany the substrate, which are used for biogas
production are constitute
of 48 % animals
excrement, 26 % organic waste and 26 % renewable
raw materials.
The biogas is end product of fermentation, which
is conbustive and it is composed as:
50-75 % methane
25-45 % carbon dioxide
2-7 % water
< 2 % oxygen
< 2 % nitrogen
< 1 % amonia
< 1 % hydrogen sulphide
The content of energy is depended on the methane
concentration. For example when a substrate as the
fats and starch are used, which are easy to break
down in fermented mass, the content of methane is
greater. Compare to fossil fuels the calorific value of
1 m3 of biogas is substitute for 0.6 l of heating oil.
TABLE III
BIOGAS BASIC INFORMATIONS [1]
Raw materials:
Annual yield per hectare
Fuel equivalent
Market price
CO2 reduction
Technical information

Maize and other energy

producing plants
4950 m3 or 3560 kg
1 kg of methane substitutes 1.4 l
of petrol
-

Biomethane can be used in

natural gas vehicles witout
adjustments

In Germany the producing of biogas is around 23-24

billion m3/year. From agricultural sector provide
about 85 % of them. When we compare the amount
of biogas to all energy consumption in Germany, it
represent only about 3 %.
The bigest disadvantages of gas sources are transport
and storage compare to liquid fuels. Methan require
more space for storage and their density of energy is
lower. For example in natural gas vehicles, the
methane is compressed in peculiar pressure tanks,
which are instaled and adapted to pressure of 200 bar.
The most benefits of using biogas compare to petrol
and diesel is very low emission of several toxic
substances, mainly nitrogen oxides or reactive
hydrocarbons, which can be reduced by up 80 %.
Biogas is able to use for runing the motor vehicles.
For example in Germany is abour 56 million cars and
only 35000 run with natural gas. These cars could use
biogas in the future without technical adjustment.

V RAPEOIL
The rapeoil is not only used as raw material for
biodiesel production, but can be used originally in
specially diesel engines. From wide variety of
vegetable oils the rapeoil is most economically
cultivated compare to sunflower or soya oil.
Production of rapeoil is divided on two groups:
one i s industrial and the second is decentralised.
Decentralized yield of rapeoil is conducted on farm
by cold pressing, where the rapeseeds are pressed by
mechanical force at maximum of 40C. After
pressing suspended solids are capture by filtration or
sedimentation from oil. The press cake contain over
than 10 % of oil, which is used later as a fodder.
Industrial production is centralized, where the rape
seeds are presed at high temperature aroun 80C and
the oil is extracted from cake with solvent. Latter the
solvent is separated from the oil by evaporation.
Next step is refining process, because the oil still has
unpleasant components. The end product is fully
refined and has quality as edible oil.
The properties of of rapeoil are specific, which
makes it different from diesel and using in
combustion engines is possible after refitting. First
large differnt is viscosity, which in low temperatures
is up to 10 times higher than diesel. High viscosity
leads to technical challenges in winter and cold
starting in conventional engines, because the
flashpoint is significantly higher that in normal
diesel. On the other hand the storage and transport of
oil is more safety. Rapeoil is not included in any
hazard classes according to the Ordinance for
Flammable Liquids.
TABLE IV
RAPEOIL BASIC INFORMATIONS [1]
Raw materials:
Annual yield per hectare
Fuel equivalent
Market price
CO2 reduction
Technical information

Rapeseed oil
1480 l/ha
1 l of rapeseed oil substitutes
0.96 l of diesel
0.55-0.75 EUR/l

>80 %
Engine refit necessary

When the raw rapeoil is required to use as a fuel, the

engine should be reffited adapt to the viscosity and
combustion properties of the rapeoil. Same refitting
bases on the pre-heating system, where the fuel and
injection system are heated before starting engine.
Other way based on 2-tank system, where engine
start with diesel and after required time is changed to
rapeoil.
In Germany is about 250 filling stations with rape or
vegetable oil. The pure fuel cost is lower than diesel,
but the costs for refitting and more frequent oil
changes are incurred, from these reasons a driving car
with rapeoil fuel becomes economically viable from
100000 kilometers of driving.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
101

VI SYNTHETIC BIOFUELS
The synthetic fuels are new dewelopment and
still not available on the market. The raw material for
this kind of synthetic fuels is biomass. The idea of
synthetic biofuels is becoming the biomass to liquid
form, which could be adapted to current engine
concepts. The most advantage is that very many
different materials, such as biological waste, straw or
wood, can be used for production the synthetic
biofuels.
The process of manufacture the synthetic fuels from
biomas is in two stages. In the first stage, which is
called gasification, a synthetic gas is produced in a
reactor, where the biomass is placed and broken
down in the presence of heat, pressure and oxygen.
After first stage the synthetic gas compose mainly of
hydrogen, carbon monoxide and carbon dioxide. In
the second stage componets are synthesised from
this, which is able to be processed to synthetic
biofuel end product.
The chemical properties of hydrocarbons in
synthetic biofuels allow efficient and complete
combustion with low exhaust gas emission. The
properties can be influenced by changing parameters
during synthesis such as the pressure, temerature and
catalysts. Modern fuels and engines could be adapted
each other and could be able to fulfil the
requirements for the less emission and improving
energy efficient.
TABLE V
SYNTHETIC BIOFUELS BASIC INFORMATIONS [1]
Raw materials:
Annual yield per hectare
Fuel equivalent
Market price
CO2 reduction
Technical information

VII CONCLUSIONS
In spite of full of promise potencials of biofuels,
the combustion engine in combination with fossil
diesel and petrol will dominate the transport sector in
the future. Propably be not the only strategy to save
or replace the fossil fuels. Some of them have
immediate effect and other like biodiesel or
bioethabol will contribute large amount of fossil
fuels. Till 2010 they will share about 5 % of fuels
market by mixing with conventional fuels.
From all of biofuels, the synthetic biofuels give
the greatest opportunity of potential quantities in the
long term, but they will make a contribution after
2010, when them production will be on an industrial
scale. Till 2020 their consumption share could
increase to 9 %.
The opportunities of biofuels consumption share is
depended on tax incentives and promotion of
research.

[1]
[2]
[3]

[4]
[5]

[Link]
Harry D. Saunders, Virtual biofuels- A cheaper, better, faster
alternative; Energy Policy 36, 1247-1250, 2008.
C. Bomb, K. McCornick, E. Deurwaarder, T. Kaberger,
Biofuels for transport in Europe: Lesson from Germany and
the UK, Energy Policy 25, 2256-2267, 2007.
J. Peters, S. Thielman, Promoting biofuels: Implications for
developing countries, Energy Policy 36, 1538-1544, 2008.
E. Kondili, J. Kaldellis, Biofuel implementation in East
Europe: Current status and future prospects, Renewable and
Sustainable Energy Reviews, 11, 2137-2151, 2007.

Plants and wood

4030 l/ha
1 l of synthetic fuel substitutes
0.97 l of diesel
-

>90 %
Can be used in pure form or in
mixtures without adjustment of
engine

Based on the current knowledge, in the future will be

possible to transfer about 50 % of energy contained
in vegetable raw materials to synthetic biofuels under
optimum conditions. From this reason it can result in
large carbon dioxide saving.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
102

Guideline for the Investigation of PEM Fuel Cell

Systems in Automotive Applications
M. Heuer, M. Kbisch, G. Heideck, Z. A. Styczynski
Otto-von-Guericke University Magdeburg
Universittsplatz 2
Magdeburg, Germany

Abstract-In the context of a research project in the field of

energy conversion and drive systems at the Otto-von-Guericke
University Magdeburg, four proton exchange membrane (PEM)
fuel cell systems of the newest generation have been put into
operation. The investigations refer to stationary and mobile
applications as well as to lifetime tests of the fuel cell. The plants
are presented and scheduled measurements plus the research
goals are described. The operational behaviour of the fuel cell and
the overall system are evaluated for long-term investigations. A
further emphasis is on the behavior of several fuel cells, which are
operated with different gas compositions. Because the
measurements are still in the early stages, the report mainly
describes the project goals and gives an overview of fuel cell
technology.

I.

INTRODUCTION

Both the conventional and the alternative energy industry are

undergoing strong changes due to the intensive price
fluctuations of fossil energy sources in the last few years [1].
The future global energy supply can be realized only by an
increase in the amount renewable energy, in order to ensure a
safe and stable energy supply in public and private life. Apart
from the renewable energy from sun, wind, earth energy and
biomass, hydrogen is another option, which through industrial
or renewable processes makes it possible to convert chemical
energy in fuel cells into electricity and use it at any time,
independent of the volatile energy supply from alternative
energy sources.
The electrical and thermal energy supply from the fuel
hydrogen in fuel cells has already been successfully tested at
research institutions and in the industry for over 20 years. The
market introduction of fuel cell systems to the home energy
supply or to the driving power in vehicles is, however, delayed.
This delay is due to the current lack of a comprehensive
hydrogen infrastructure and also because material problems
with the electrolyte and the membrane of the fuel cell have not
been solved yet.
In a research project in the field of automotives in a scientific
instrumentation proposal of the German Research Foundation,
four PEM fuel cell systems with an electrical power between
400W and 5000W are examined under different aspects and
future areas of application of PEM fuel cells are tested. The
measurements will begin fully in the spring of 2009.

II.

INTRODUCING OF FUEL CELL SYSTEMS

The investigations are directed at four PEM fuel cell systems

in the electric power range between 400 and 5000W. These are
low-temperature fuel cells with a maximum operating
temperature of 80C on the primary water side. The high
temperature PEM fuel cells work at temperatures of up to
180C and have been tested only for a few years in the field of
research [2]. The plants in this project, however, are not
designed for this type of high temperature PEM fuel cells.
The fuel cell group consists of three systems with an
electrical output in the range of 400-500W and one system with
an output power of approximately 5kW of electric and 10kW
of thermal power. Each plant is different in its procedural
process structure and is used for various measurements and
investigations. The three small plants are named the
Reformate, Hydrogen and Experimental plants.
The reformate fuel cell system is operated with hydrogen
from an electrolysis plant and carbon dioxide from pressure
cylinders. The fuel composition is derived from the actual
reforming of natural gas or biogas after the reforming stage.
The real proportions of hydrogen and carbon dioxide through
the reformation were measured by another fuel cell system
with an integrated reformer. These measurements are now
available for the simulated gas composition of the reformate
plant. Depending on the efficiency of the reformer and electric
power of the fuel cell, the proportions of hydrogen is at a
maximum of 70% and carbon dioxide at 30%. This unit is
designed for long-term studies in the stationary operation for
the domestic energy supply. Both the fuel cell system as well
as the components of the plant will be investigated after several
thousand hours of operation. The hydrogen plant was
developed with much more variability in its operation mode
and operates with pure hydrogen from the electrolysis plant.
Different modes of operation can be implemented, which can
be used for different applications. For example a recirculation
of unused hydrogen, which does not react at the membrane, is
integrated by being pumped back to the inlet hydrogen channel
of the fuel cell. The practical effect of recirculation on the fuel
consumption and material compatibility of the pumps on the
humid and heated hydrogen are an essential aspect of the
investigations of this facility.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
103

Figure 1. Fuel cell test benchesReformate,Hydrogenand Experimental

(from right).

Another focus is the behaviour of the bipolar plates under

various operating conditions. The bipolar plates form the gas
channels through which the fuel and oxygen lead to the
membrane, whereby hydrogen is on the anode side and oxygen
of the ambient air is on the cathode side. The correct
construction of these channels is responsible for low pressure
drops and for an effective use of the air compressor. The air
compressor has the largest energy consumption of the fuel cell
system. The bipolar plates have been especially designed for
each fuel cell system by the manufacturers. The test results of
this new generation of bipolar plates and the use of a new
membrane show development progress in this field. The
hydrogen plant is suitable for investigations in the mobile and
portable fields in which pure hydrogen and no gas
compositions are used. The system will be tested in the range
of auxiliary power units (APU) for automobiles and the
possibilities will be explored for a reduction of fuel
consumption through the use of a fuel cell.
The third plant is used for system optimization. Individual
components are tested and developed with the goal of
minimizing energy expenditure and simplifying and optimizing
the management processes. The development work in this area
should be introduced in the other plants, so that the entire
system is improved in its efficiency.
The 5kW plant will be supplied with pure hydrogen from a
tank reserve on the university campus. The main focus of this
plant lies in the mobile application. The various modes of
hydrogen management in the system and the ability of the
moistening of the reaction gases to the optimal energy
conversion at the cell membrane are the subject of
investigations.
All four units are connected via a network and can be
controlled in the laboratory as well as with remote access.

Figure 2. 5kW fuel cell facility for mobile tests.

III. FUEL CELL RESEARCH FOR STATIONARY APPLICATION

The power supply for households is being decided ever more
strongly by decentralized energy installations directly at the
site of the end customers. Solar systems, small wind turbines or
wood pellet heating systems will find increased use in single
and multi-family houses. In this application market, fuel cell
systems offers a great potential as combined heat and power
plants, to provide both the electrical and thermal energy. The
prerequisites include 40,000 operating hours, low downtime,
comparable costs in the range of existing home power systems,
and the useful hydrogen production from renewable energy
sources. However, an introduction to the market of such
stationary fuel cell plants in private households is expected yet.
At the moment different projects involving fuel cell plants are
being used and tested in households by the end customers in
their everyday employment [3].
The tests for stationary applications are based on load profile
curves of single and multi-family houses. Load profiles are
copied during a longer period. Here, the actual performance of
the fuel cell is scaled down to the power in the household.
From the result of several thousand operation hours, the
achievement performance of the fuel cell and life time
prognoses should be determined. Figure 3 shows the load
profile of a single family house on one summer day at the
weekend. From the existing data records, the fuel cell system
can now be operated as a house power supply plant.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
104

Saturday 02.08.2003
7
6

Power [kW]

5
4
3
2
1

00
:0
01 0
:0
02 0
:0
03 0
:0
04 0
:0
05 0
:0
06 0
:0
07 0
:0
08 0
:0
09 0
:0
10 0
:0
11 0
:0
12 0
:0
13 0
:0
14 0
:0
15 0
:0
16 0
:0
17 0
:0
18 0
:0
19 0
:0
20 0
:0
21 0
:0
22 0
:0
23 0
:0
00 0
:0
0

Figure 3. Load profile of a four person household [4].

IV. FUEL CELL RESEARCH FOR MOBILE APPLICATION

The automobile and its different propulsion principles are
undergoing great changes. The European Union calls for
drastic reductions of emissions in the transport sector through
new innovative drive concepts to ensure mobility and to be
independent of limited fossil fuels. This can be achieved by
more efficient engines, the integration of energy storage such
as batteries to minimize the consumption and to reduce local
emissions in hybrid vehicles as well as the long-term market
conquest by fuel cell and electric vehicles. The battery
technology of electric vehicles is not currently in a position to
travel distances in excess of 100km, so this energy storage is
used especially in urban areas. The fuel cell can travel
distances of up to 300km using new storage technologies, like
the 700 bar storage of gaseous hydrogen [6]. For example, fuel
cell buses have already been driving in the public traffic for
several years in Hamburg and Berlin.
Illustration 4 shows the energy expenditure for a distance of
100km for different engines, whose fuel originates from
different energy resources. It can be clearly seen that fuel cell
vehicles possess less energy expenditure than conventionally
moved vehicles only during renewable hydrogen production.
12
1,
5
95
,4

46
,8

18
,9
19
,8
0,
9

8,
1

8,
1

20

20
,7
26
,1

26
,1

40

18
,9

35
,1

49
,5

54

57
,6

60
,3

80

52
,2

energy input in kWh/ 100km

120

60

0
engine

fuel

process chain

energy resource

otto engine
(DISI)

diesel engine
(DICI)

benzine

diesel

production/
production/
transportation/ transportation/
refining
refining
oil

oil

electric engine electric engine electric engine electric engine

(fuel cell)
(Li-Ion-battery)
(fuel cell)
(Li-Ion-battery)
pressured
hydrogen

electricity

pressured
hydrogen

electricity

power
generation/
power gird/

power
generation/
power gird/

power
generation/
power gird/

power
generation/
power gird/

wind/ water/
sun

wind/ water/
sun

Eu-energy-mix Eu-energy-mix

producer chain

The reason for this is that at the moment, the production of

hydrogen is still energy-intensive. Battery vehicles exhibit in
contrast smaller energy expenditure in the European energy
mix and especially with regard to the renewable energy supply.
As previously stated, the 5kW system is used in the research
project for mobile applications. Figure 5 shows the
experimental scheme between the fuel cell and an electric
motor. The energy of the fuel cell is delivered to a load and a
scaling calculator increases the electric power of 5kW up to
20kW. The use of power electronics, inverters, electrical motor
and storage systems such as batteries and super capacitors
constitute the entire power train of a vehicle. The whole system
is tested by driving cycles and the interaction of the different
energy storages is investigated. During rapid load changes, the
super-capacitors should guarantee the dynamics of energy
supply.
V. LIFETIME PREDICTION OF FUEL CELL AND GREEN FUEL

140

100

Figure 5. Scheme of the mobile fuel cell system [7].

vehicle

total

Figure 4. Energy expenditure of different engines and fuels [5].

The lifetime of the fuel cell stack is the main criterion for the
market launch of fuel cell systems for various applications.
Also after 5,000 or 40,000 operation hours the electrical power
output may decrease only insignificantly. The lifetime
estimation can be accelerated by lifetime tests. This requires
more stringent operating conditions. It should be noted,
however, that the meaningfulness of such accelerated tests is
limited, because key operating conditions such as downtime
and its impact on the fuel cell will be not considered. In
contrast, real operating conditions of several plants under
realistic dynamic loads are more meaningful. A disadvantage is
however the expenditure of time.
A combination of accelerated tests and the influence of key
parameters and criteria that can only occur under real
conditions will be introduced in the investigations. From these
measurements lifetime curves arise for one type of fuel cell,

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
105

depending on the design of bipolar plates and type of

membrane used and other components. From the comparison
of several fuel cells then prognoses for the lifetime and for
influence parameters are to be identified.
A further influence parameter on the lifetime is the quality of
the fuel and its composition. By using different gas
compositions and humidity ratios influencing factors have to be
identified and have to be integrated in lifetime forecasts. Above
all, the quality of the fuel gas and its monitoring is an
important aspect in these studies.
VI.

CONCLUSION

This paper gives an overview of the future studies on four

new PEM fuel cell systems for mobile and stationary
applications. In addition to lifetime predictions and
measurements in simulated fuel compositions, system
optimization is also in the foreground. The results feed directly
into the system optimization of fuel cell systems of project
partners.

REFERENCES
[1] HZwei, Magazine for hydrogen and fuel cells, Kremmen: October 2008,
pp. 3.
[2]

J. Zhang et al. High temperature PEM fuel cells, Journal of power

[3]

A. Ballhausen, Small stationary CHP-systems- overview about the status

[4]

C. Heyde, Energy management strategies for grid connected

sources, 2006.
quo, fuel cell conference, Braunschweig: 2008, pp. 115-124.
photovoltaic plants with battery storages, diploma thesis, Otto-vonGuericke University, Magdeburg: 2006.
[5]

"Well-to-Wheels analysis of future automotive fuels and powertrains in

the european context, study of CONCAWE, EUCAR und JRC, 2005.

[6] HZwei, Magazine for hydrogen and fuel cells, Kremmen: July 2008, pp.
12-14.
[7] M. Kbisch, Test Bench Development for Investigation Fuel Cell
Technology in Automotive Application, Proceedings of the ReDiPS
Workshop, Athens: September 2007, pp.66-71.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
106

Wind Turbine Driven Doubly-Fed Induction

Generator with Grid Disconnection
[Link] Babu #1, [Link]#2 and [Link]*3
#

Department of Electrical Engineering, National Institute of Technology Rourkela

Rourkela-769008 (India)
1
2

bcbabunitrkl@[Link]
kbmohanty@[Link]

* Department of Electrical Engineering, Indian Institute of Technology Madras

Chennai-600036 (India)
3

ee08d002@[Link]

AbstractThis paper describes the transient behaviour of a

doubly-fed induction generator (DFIG) driven by wind turbine
after its disconnection from the grid. The induction machine runs
at a specific speed with the stator disconnected from the grid
(Is=0), the rotor is suddenly excited with slip-frequency voltages
derived from voltage regulators so as to produce commended
open-circuit stator terminal voltage. Behaviour under varying
rotor speed typically observed in wind turbines is also reported.
A MATLAB computer simulation study was undertaken and
results on 1.5 kW wind turbine are presented.

II. INTRODUCTION
For stand alone or autonomous operation, mostly single
induction generator or parallel operated induction generators
are focused according to available analysed references. These
induction generator driven by the individual prime movers
employed excitation capacitor bank to build up desired
voltage via self-excited phenomena. Hence the value of the
excitation capacitor bank and the rotor speed determine the
magnitude of the generated voltage and its frequency. Both
Index Terms- Doubly-Fed Induction Generator (DFIG), Variable- voltage and frequency need to be controlled to feed the power
speed wind turbine, Dynamic Modeling, Grid Disconnection, to the load. But for grid connected operation, there are two
Transient Analysis
types of generators are used. (i.e., single output and double
outputs).In order to feed the active power to the grid, the
machine should run at a speed greater than the synchronous
I. NOMENCLATURE
speed of the revolving magnetic field. (i.e. slip should be
Vqs, Vds are the three-Phase supply voltages in d-q reference negative).The single output generator feeds active power to
the grid via only stator side and double output generator feeds
frame, respectively
iqs ,ids are the three-Phase stator currents in d-q reference electrical power to the grid via both stator as well as rotor side.
The latter is also called static Kramer, double-fed or double
frame, respectively
qs ,ds are the three-Phase stator flux linkages in d-q reference outputs induction generators. This is only the generator which
generates the power more than rated power without
frame, respectively
Vqr, Vdr are the three-Phase rotor voltages in d-q reference overheating. Besides, this kind of power generation usually
causes problems in the utility grid system. Because the control
frame, respectively
iqr ,idr are the three-Phase rotor voltages in d-q reference frame, on active and reactive power of the machine is complex one.
Wind turbines often do not take part in voltage and frequency
respectively
qr, dr are the three-Phase rotor voltages in d-q reference control and if a disturbance occurs, the wind turbines are
disconnected and reconnected when normal operation has
frame, respectively
rs ,rr are the stator and rotor resistances of machine per phase, been resumed. As the wind power penetration continually
increases, power utilities concerns are shifting focus from the
respectively
Lls, Llr are the leakage inductances of stator and rotor windings, power quality issue to the stability problem caused by the
wind power connection. In such cases, it becomes important
respectively
e, r are the supply and rotor angular frequency (electrical to consider the wind power impact properly in the power
system planning and operation. This paper will focus on the
speed), respectively
grid-connected induction generator feeding power with DOIG
Te is the electromagnetic torque
Ps, Qs are the stator-side active and reactive powers, during steady state and transient conditions.
This paper describes the transient behaviour of a doublyrespectively
fed induction generator (DFIG) driven by wind turbine after
P is the Number of poles
its disconnection from the grid. The induction machine runs at
a specific speed with the stator disconnected from the grid
(Is=0), the rotor is suddenly excited with slip-frequency
voltages derived from voltage regulators so as to produce

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
107

commended open-circuit stator terminal voltage. Behaviour

under varying rotor speed typically observed in wind turbine
is also reported. A MATLAB computer simulation study was
undertaken and results on 1.5 kW wind turbine are presented.
III. DFIG DYNAMIC MODELLING
A commonly used model for induction generator
converting power from the wind to serve the electric grid is
shown in [Link] stator of the wound rotor induction
machine is connected to the low voltage balanced three-phase
grid and the rotor side is fed via the back-to-back IGBT
voltage-source inverters with a common DC bus. The network
side converter controls the power flow between the DC bus
and the AC side and allows the system to be operated in subsynchronous and super synchronous speed. The proper rotor
excitation is provided by the machine side power converter.

Fig. 1. Model of DFIG Wind Turbine

The general model for wound rotor induction machine is

similar to any fixed-speed induction generator as follows.
A. Stator Voltage Equations

Vqs = pqs+ds + rsiqs

Vds = pdsqs + rsids

(1)

E. Stator Flux Linkage Equations

qs = (Lls + Lm )iqs + Lmiqr

ds = (Lls + Lm )ids + Lmidr

F. Rotor Flux Linkage Equations

qr = (Llr + Lm )iqr + Lmiqs

dr = (Lls + Lm )idr + Lmids

Vdr = p dr ( r )qr + rr idr

(2)

C. Power Equations

3
(Vdsids + Vqsiqs )
2
3
Qs = (Vqs ids Vds iqs )
2
Ps =

(3)

D. Torque Equation

Te =

3P
(ds iqs qs ids )
2 2

(4)

(6)

IV. TRANSIENT ANALYSIS DURING GRID DISCONNECTION

Fig (2) shows the three phase stator voltages under normal
operating conditions. When the induction machine is running
at a particular speed while the stator disconnected from the
grid. So the rotor is suddenly got excited due to slip frequency
rotor voltages from the voltage regulators in order to produce
the commended stator terminal voltage. Since the variation of
speed of the rotor, torque could also be varied on the machine.
Fig (3) shows the transient response of the stator voltage of
induction generator under torque disturbance. It is found that
the voltage of the stator becomes slightly small value after
disturbance. Fig (4) shows the transient response of the rotor
voltage of induction generator under torque disturbance. Fig.
(5) shows the transient response of the active power of the
induction generator during disconnection. When induction
generator is disconnected from the grid, the active powers
supplied from induction generator decreases and quickly
recover to original value after re-closed to the grid. The
changes in a reactive power are also shown in Fig. (6). It is
observed that the reactive power absorbed by the induction is
also decreases rapidly, but the part of reactive power would be
supplied by rotor side converter for compensation during reclosed to the grid. Negative values of active and reactive
power indicate the machine working in generating mode.

B. Rotor Voltage Equations

Vqr = p qr +( r )dr + rr iqr

(5)

Fig. 2. Response of Three phase Stator voltages of DFIG

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
108

Fig. 3. Transient Response of Stator voltage During Grid Disconnection

Fig. 4. Transient Response of Rotor voltage During Grid Disconnection

Fig. 6. Transient Response of Reactive power During Grid Disconnection

IV. CONCLUSION
In this paper, dynamic characteristics of double-fed
induction generator has been studied during abnormal
conditions of the grid. For this, dynamic d-q model was used
to derive the dynamic equations of such machine in a
synchronous reference frame. The choice of synchronous
rotating reference frame makes it particularly favourable for
the simulation of double-output configuration in transient
conditions. When the stator is disconnected from the grid, the
rotor is suddenly got excited due to slip frequency rotor
voltages from the voltage regulators in order to produce the
commended stator terminal voltage. So active and reactive
power of the machine have been decreasing rapidly. For
reactive power compensation during these conditions, rotor
side converter has to supply necessary reactive power.
REFERENCES

Fig. 5. Transient Response of Active power During Grid Disconnection

[1] [Link], [Link] and J.X. Saenz, Modeling and control of a wind
turbine driven doubly fed induction generator, IEEE Energy
Conversion, vol. 18, pp. 194204, June. 2003.
[2] Yazhou Lei, Alan Mullane, Gordon Lightbody, and Robert Yacamini,
Modeling of the Wind Turbine With a Doubly Fed Induction
Generator for Grid Integration Studies, IEEE Energy Conversion, vol.
21, pp. 257264, Mar. 2006.
[3] C. S. Demoulias and P. S. Dokopoulos, Transient behaviour and self
excitation of wind-driven induction generator after its disconnection
from the power grid, IEEE Energy Conversion, vol. 5, pp. 272278,
1990.
[4] P. S. Nagendra Rao and S. S. Murthy, Performance analysis of grid
connected induction generators driven by hydra/wind turbines
including grid abnormalities, in Proc. 24th Intersociety on Energy
Conversion Engineering Conference, 1989, pp. 2045-2050.
[5] L. Wang, Ya-Feng Yang and Sung-Chun Kuo, Analysis of Gridconnected Induction Generators Under Three-phase Balanced
Conditions, in Proceedings of IEEE 2002, pp. 413-417.
[6] [Link], H. Polinder, and [Link]. Dynamic modeling of
wind turbine with doubly fed induction generator, in Proc. IEEE
Power Eng. Soc. Summer meeting, Vancouver, BC, Canada, Jul. 15-19,
2001.
[7] [Link], Srinivas Chellapilla, Doubly-fed induction generator
for variable speed wind power generation Transactions on Electric
Power System Research, Vol. 76, Jan 2006. pp. [Link],
[Link], Wind Electrical Systems, Oxford University Press.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
109

Possibilities of Utilization of Energy Briquettes

Gabriel Borowski

Lublin University of Technology

20-618 Lublin, Nadbystrzycka 38, POLAND
e-mail: [Link]@[Link]
Abstract. The analysis of mineral coal briquettes from charcoal
and biomass together with molasses as a binding agent for
utilization in industrial energy were presented. The physicalchemical properties of the material investigated, consisting of raw
waste material, are described. The process for preparing the
material into briquettes comprised of crumbled and mixed
constituents, together with the specific reduction of moisture of the
mixture, are presented. The effect of the selected parameters of the
consolidation process in the pressing stamp on the toughness of
briquettes is investigated. Analysis was made of such parameters
as: contribution of the biomass in the mixture of charcoal and
binder, moisture in the mixture, press pressure on the formation,
and seasoning. The results confirm the possibility of utilizing high
quality energy briquettes. They possess the high material strength
as well as high value combustible fuel, which qualifies them for
utilization in industrial energy.

1. INTRODUCTION
Coal-biomass briquettes are an ecological source of energy
and even frequently utilized in the production process for
heating energy as well as in electric heating, such as in
individual households. Briquettes produced at present contain
pit coal, brown coal (lignite), coke together with biomass in the
form of sawdust, peat, straw or other raw materials. Briquettes
thus produced are without any additions or additional binding
agents. Frequently used binders are starch, sugar, soda lye,
water glass, lime, gypsum, and many others. The addition of
these fine-grained binding agents to the composition of the
briquettes increases their toughness [1, 2, 3].
For making the briquette charcoal-biomass mixture,
stamping presses as well as rollers may be used. Rolled
briquettes are definitely more efficient than those that are
punched, and ought to have a wider application for production
continuity. Stamping presses, however, check the determined
properties of briquettes in laboratory conditions and also for
the production of the specimens in small series. The mixture
for merging on the press has to be appropriately prepared. Each
individual ingredient has to be added in determined
proportions, binder added, then thoroughly mixed together and
the paste dried to a certain moisture [3].
Coal-biomass briquettes intended for burning in furnaces are
characterized by a high energy effect. Their material strength
has to be adjusted to transport conditions and storage, through
which the toughness ought to be kept throughout the seasonal
periods. In the case of transportation in sealed containers, the
briquettes should be marked for additionally high water
resistance.

Briquettes with biomass used as energy fuel comply with

the standard regulations for the protection of the environment,
especially in area of low emissions of harmful substances into
the atmosphere [7]. The mass production of briquettes
requires, moreover, preparation of high quality criteria and
repeatable quality of this product. Appropriate laboratory
experiments were carried out to put this aim into effect. The
present work recommends preparation in this way of the
mixture and manufacture of coal-biomass briquettes on a
laboratory scale.
2. MATERIALS fOR ENERGY BRIQUETTES
The materials investigated were pit coal and sawdust from
deciduous trees (oak and beech). The binder added to the
mixture was molasses, a by-product obtained during the
refining of sugar. The selection of this binder from among many
others was motivated by the results of the authors investigations
with fine-grained briquette materials which confirmed the effect
of molasses on the material strength of briquettes [3]. Also
significant is the fact that it did not create any increase in
harmful emissions during the burning process of briquettes of
molasses in comparison with test controls.
The value of burnt biomass of deciduous wood rose in
median to 19,000 kJ/kg. In comparison, the value of burning
pit coal rose in median to 28,000 kJ/kg (Table 1). During the
burning time the biomass produced a small amount of ash
which did not contain harmful substances and may be utilized
as mineral fertilizer [6].
The basic chemical ingredients of pit coal and biomass
employed for energy are the same. The individual elements in
chemical conjunction appear, however, in different proportions. The resulting measurements presented in Table 1
confirm that the biomass contains about four times more
oxygen, twice as few charcoal elements, and likewise sulphur
and nitrogen. The consequence of these properties is a higher
air content and higher reactivity of biomass [8].
From the ecological point of view, the advantageous
characteristics of wood biomass are significantly lower, in
comparison with coal, sulphur and ash value measurements.
The contribution of ash formed in wood burning is on the
level of 2% and is significantly lower than for burning coal.
The value of the air part, however, is significantly higher in
biomass than with coal. The content of the element coal in the
biomass is on the level typical for organic combinations [5].
Biomass possess a lower pouring density and therefore a
bigger ingredient surface is required in comparison to the
coal storage area. The biomass density increases on average

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
110

TABLE 1
CHARACTERISTIC OF CHARCOAL AND BIOMASS
Parameter

Pit coal

Density poured (kg/m3)

Combustible value (kJ/kg)
Moisture value in investigation (%)
Value of air part (%)
Ash value (%)
Sulphur value (%)
Oxygen (%)
Coal (%)
Nitrate (%)

880
28,000
7.1
26.6
12.2
0.9
7.0
81.0
1.1

Biomass of
deciduous wood
80
19,000
12.0
66.5
2.0
0.1
30.2
42.0
0.5

80 kg/m3 at the same time as coal increases to ca. 880 kg/m3

(Table 1).
The characteristic biomass attributes in its fresh form is a
high value of moisture which increases as high as 50%. It is
advantageous to decrease its moisture to a value lower than
15%. A high moisture content in the biomass also has an
influence on its cost and profit. A higher weight of biomass
relative to the appreciable water content increases the cost of
transportation. The utilization of waste wood is therefore
profitable through application of drying and integration of raw
materials before its transportation to energy works. A very
important characteristic mixture of coal-biomass is the
complete additiveness of organic substances of both fuel. In
this mixture, the coal plays a stabilizing role in the burning
process. In comparing energy, 2 tons of biomass are equal to 1
1.5 tons of pit coal [4].
The combustible value of coal briquettes with 20% massive
participation of wood biomass rises to ca. 24,000 kJ/kg, which
is sufficient in order to employ it as a source of warming
energy in electric generating plants. Such briquettes, moreover,
are an ecological fuel complying with regulations for
protecting the environment, relative mainly to the limitation of
SO2 emissions into the atmosphere, thanks to which
desulphurisation of fumes is not required [9].
3. TOUGHNESS OF BRIQUETTES
Fine grain coal belongs to the materials that showed weak
susceptibility for integration in pressing into briquettes. It is
therefore desirable that the activity has the aim of changing the
properties of the material. In investigations to date, the
integration of fine grain materials has played a significant role
factor combining the prepared material [2, 3].
The preparation of material for making into briquettes
includes the crumbling and mixing as well as drying to a
moisture level below of 12%. An electric mill is used for the
crumbling. The relatively significant difference in the
physical properties of coal and fibrous materials suit to two
types of mill coal hammer and shears (cutting) for fine
graining into biomass. Coal together with sawdust were
crumbled to the moment granules of less than 2 mm are
obtained for analytical tests. The prepared briquettes were
investigated for the balance participation of the biomass
decreased in relation to coal to 20% and 25%. In

accomplishing, the briquettes checked also only the coal. The

molasses content in the role of binder, defined previous
investigations, decreased to ca. 8% [3].
The ingredients were mixed in an electric mixer and then
taken for thermal drying to a precise level of moisture.
Investigation of the moisture was accomplished with the aid of
laboratory moisture tester from the KETT firm, type FD-620.
The drying temperature was increased to 100C. The prepared
briquette mixture was placed in the laboratory press. The
maximum pressure of each individual press was increased to
35 MPa, whereas the press displacement was 200 mm.
Preparation and accomplishment of special forming unit to
briquetting of fine grain waste materials were assembled in the
matrix and punch. The matrix enables obtain briquettes of
cylindrical shape and a suitable height of briquettes obtained of
approximately 2/3 diameter. It is possible to quickly change the
matrix and stamp for the purpose of obtaining briquettes of
different dimensions and form.
The toughness of the briquettes were defined through
resistance to gravity drop, together with the strength of
compression. Resistance of the briquettes to gravity drop was
evaluated through the percentage of decrease in mass after
dropping the briquettes at least three times from a height of 2.0
m on to a steel plate. After each drop, the test piece was put
through a sieve with openings smaller than the minimum
dimension of the briquette. Briquette resistance to gravity is
illustrated in the equation below:

Bz
100 %
B
where: B briquette mass before dropping,
Bz mass remaining in filled sieve.
K=

(1)

The gravitational resistance to drop should be attain a

value higher than 90%. The value of compression strength
destroying the briquette was defined experimentally as
follows, that the cylindrical disc was placed between the flat
surface of the compress experimental piece until the moment
of destruction of its structure. The investigation carried out
used an endurance test machine ZWICK Z100. The value of
compression strength destroying the briquette should be attain
at least 1.5 MPa.
4. ANALYSIS OF RESULT OF RESEARCH
Fresh and seasoned briquettes were tested. Analysis of the
following parameters was undertaken: biomass contribution,
moisture of the coal-biomass mixture with binder, pressure of
stamping press and seasoning. Data from the analysis is
presented in Table 2, as well in Figures 1 and 2.
The data presented in Table 2 shows that the majority of
examples of gravitational resistance to drop obtained a value
higher than 90%. This resistance is insufficient through
integration in the material of moisture greater than 10%.
It was ascertained that the mixture should contain a 6 9%
of moisture. High resistance of the briquettes obtained by a

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
111

TABLE 2
TOUGHNESS OF SELECTED COAL-BIOMASS BRIQUETTES

Test Moisture
No.
(%)

1
2
3
4
5
6
7
8

11.5
9.6
8.0
7.9
7.5
7.4
6.6
6.3

Punch
pressure
(Mpa)

Biomass
participation
(%)

30.0
30.0
35.0
35.0
30.0
35.0
35.0
30.0

20.0
20.0
20.0
25.0
25.0
25.0
0.0
20.0

Maximum
compression
Resistance
strength
to gravity drop
destroying
(%)
briquettes
(Mpa)
fresh seasoned fresh seasoned
1.08
4.81
4.42
5.30
6.03
7.45
1.19
4.84

1.54
7.37
6.80
8.14
9.24
11.64
1.83
7.44

80.0
89.0
97.5
95.6
94.8
97.1
82.6
93.7

82.0
90.5
98.2
96.3
95.6
99.0
83.5
94.0

punch press with pressure of 30 as well as 35 MPa. through

which the greatest pressure value applied the biggest
degree of crush; therefore, in effect, the material strength
was higher. This is significant if the contribution of biomass
in the briquettes is increased. For example, a 25%
contribution of biomass in the samples produced with a
smaller resistance; on the other hand, however, through
increasing the strength value of pressure, samples are
produced with a high toughness.
It was observed that in gravitational resistance drop tests the
greatest resistance was proved in tests falling on to the surface
of the circular disc, and the lowest when dropped on to the
edge or flat part. This confirms the justification for
constructing such a matrix and punch, so that the briquettes had
not strong edges and even walls. Round-shaped briquettes can
be formed in a roll press.
The data presented shows that briquettes consisting only of
coal have a decidedly lower resistance to compression than
briquettes with additional biomass. The use of biomass
therefore increases the material strength of briquettes. Analysis
of the data shows that the maximum strength value of
compression is somewhat higher for briquettes with a 20%
biomass, and lower for those with 25% biomass through using
the same pressure strength in the press. It is justified to use a
higher press pressure in the example of compressing material
containing a higher biomass content.
Figure 1 presents data showing the seasoned precept of
briquettes. Seasoned briquettes are characterised by a higher
resistance to pressure than fresh ones. The results of
experiments show that in the period of three weeks
their resistance to pressure increases; however, seasoning
for a longer time is disadvantageous. There is a distinct
decrease in resistance to pressure after 50 days of seasoning
the briquettes.
Analysis of the running curve in Figure 2 shows three
phases in the compression of the briquettes. In the first phase,
a small rise in compression strength induces a big relationship
between the moving parts of the endurance machine, between

Figure 1. Reliability on compression of briquettes in relation to seasoning

Figure 2. Course of compression strength of selected briquettes : 1 made with

punch pressure of 35 MPa, 2 made with punch pressure of 30 MPa, 3 like
(2) with lower humidity of the mixture

which there are the test piece. The cause of this is a porosity of
the briquettes. In this phase the compression process is
introduced; therefore, there is a reciprocal close relationship
between the ingredients of the fine grain structure of the
briquettes. The next phase of the process is characterized by a
stepwise increase in strength up to the maximum value
whereby the test piece are destroyed.
5. CONCLUSION
The production of energy briquettes from coal and biomass
produced a high toughness. An adhesive like molasses added in
to the residual mixture produces increased material strength.
Additional increase of mass participation of biomass does not
produce a significant decrease in resistance of the briquettes,
but lessen the energy effect during their burning in furnaces.
The acquisition of a high value of resistance to gravitational
drop, together with resistance to the compression, to a large
degree depends on the moisture of the mixture, pressing stamp
pressures, and likewise the period of briquette seasoning.
It was proved that briquettes showed sufficient resistance, even
in fresh form; however, it is advantageous to use them during a
period of 24 weeks.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
112

On the basis of the results of the investigation, the following

are proposed:
1. During the process for preparing the mixture for making into
briquettes, it is essential to select definite proportions of
ingredients. The most advantageous conditions to assure the
addition of the molasses in the amount of ca. 8% for the making
of briquettes of grained materials, together with the drying of the
mixture to a moisture level within the range of 69%. Seasoned
briquettes influence their toughness.
2. The use of molasses as a binder significantly influences the
increase in the material strength. They have sufficient
resistance, even through the increase in biomass
participation, but in this care it depends on the minimal
burning output value of the briquettes in relation to the
required energy effect acquired during burning in furnaces.
3. The energy effect during burning of briquettes in industrial
furnaces should be obtained.
4. Briquetting of energy raw materials brings advantages for
decreasing the consumption of materials, together with a
lowering cost of transports and storage.

REFERENCES
[1]

[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]

Altun N.E., Hicyilmaz C. and Kk M.V., 2001. Effect of Different

Binders on the Combustion Properties of Lignite: Part I. Effect on
thermal properties. Journal of Thermal Analysis and Calorimetry.
Akadmiai Kiad with Springer Science & Business Media B.V.,
Formerly Kluwer Academic Publishers B.V., Vol. 65, No 3, 787795.
Borowski G., 2007. The possibility of utilizing coal briquettes with a
biomass. Environment Protection Engineering. Wroclaw University of
Technology, Vol. 33, No 2, 7986.
Borowski G. and Kuczmaszewski J., 2005. Investigation on Briquetting of
Metal Wastes From Bearing Industry. Waste Management & Research. Sage
Publications Ltd. Copenhagen, 5(23), 473478.
Gnutek M., Moron W. and Rybak W., 2002. Problems of burning and coburning of coal and biomass. Czysta Energia, 9, 4754.
Kubica K., 2003. Burning biomass and its co-burning with coal
technicalities, uses and limits, In: Clean and Green Energy clean air in the
Silesian Voivodship. Katowice, pp. 95106.
Niedziolka I. and Zuchniarz A., 2006. Analysis of energy from selected
types of biomass of vegetable origin. Motrol, 8A, 232237.
Purohit P., Tripathi A.K., Kandpal T.C., 2006. Energetics of coal
substitution by briquettes of agricultural residues. Energy, Elsevier Ltd.,
31, 13211331.
Sciazko M. and Zielinski H., 2003. Thermo-chemical processing of coal
and biomass. Polish Academy of Sciences, Zabrze-Krakow, pp. 125132.
Sciazko M., Zuwala J. and Pronobis M., 2006. Advantages and
disadvantages of co-burning biomass in energy furnaces in the light of
exploitation experiences during the first year of co-burning biomass on an
industrial scale. Energetyka, 3, 207220.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
113

Optimum location and sizing of passive filters in

distribution networks using genetic algorithm
M. Ghiasi, V. Rashtchi, H. Hoseini
Engineering student and faculties of Zanjan University
Zanjan, Iran
Abstract-The harmonic distortion of voltage has became an
important subject in power quality, especially after use of power
electronic equipment and nonlinear loads. Simple method to
limitation of harmonic distortion is using of passive filters. The
objective of this paper is to determine the location and size of
passive filters in distribution networks economically, by using
genetic algorithm. With using of genetic algorithm and new
coding here some of other methods limits for passive filters
locations are removed. The purpose of the genetic algorithm is to
minimize the cost of passive filters and, at the same time, to reach
the harmonic limitations defined by standard IEEE-519. This
algorithm is applied to IEEE 13-bus distribution network and the
results are shown, finally

I.

INTRODUCTION

Whit increasing use of power electronic equipment and

nonlinear loads, the level of voltage harmonic distortion in
distribution networks are significantly increasing. As
harmonics propagate through the system, they result in
increased losses and possible equipment loss-of-life. Also
Overcurrents or overvoltages resulting from resonances can
damage equipment. Additionally, harmonics can interfere with
control, communication, and protective equipment. With this
reasons different research institutes have studied about the
proper limitation of harmonic disturbance levels and released
different standards.
Among the several methods used to reduce these harmonic
disturbances, the more employed are the tuned passive filters
due to their simplicity and economical cost.
The important problem of using passive filters are
determining location and sizing of them, which is reach
standard levels of harmonic distortion with applying minimum
cost of passive filters [1, 2].
In reference [3], the objective is to propose a new approach
for designing a passive LC filter of the full-bridge rectifier by
using genetic algorithms (GAs). The objective of the fitness
function in the GA program is to find out the maximum PF of
the ac mains with the smallest inductor value.
Reference [4] describes the application of genetic algorithm
to calculate the R-L-C parameters of passive harmonic filters
that may be installed in the customers houses. The main goal
is to minimize the harmonic impedance in some specific
frequencies and to maximize the fundamental frequency
impedance in order to minimize losses.
Reference [5] presents a method that applies the
Combinatory optimization by microgenetic algorithms for the
location and optimum project of passive harmonic filters. The

minimization of total voltage harmonic distortion and active

power losses has been used as objective function. In this
reference, the problem consists of planning two passive
harmonic filters, whose harmonic tuning orders are 4.7 and 6.6,
respectively.
Reference [6] describes the application of genetic and a
microgenetic algorithm to the problem of locating and sizing
passive filters in an electric network affected by harmonic
disturbances. In this work, the objective function to be
minimized is the value of voltage harmonic distortion. In this
reference, the input data are the number of filters and the
relevant order of these filters.
In this work by using genetic algorithm and new coding,
location and sizing of passive filters is determined so with
installing them the standard levels for voltage distortion is
achieved with minimum cost of passive filters. By comparing
with other methods the number and harmonic order of passive
filters is automatically determined in this work.
Standard limitation that is considered here is standard IEEE519 for harmonic limitation.
In the next part of this paper firs modeling of passive filters,
based of new coding and finally the result of applying the new
method to IEEE- 13-bus distribution network will be illustrated.
II.

PASSIVE FILTERS

The most common type of passive filter is the single-tuned

notch filter. This is the most economical type and is
frequently sufficient for the application. The notch filter is
series-tuned to present low impedance to a particular harmonic
current and is connected in shunt with the power system
(usually near the loads that produce harmonics). Thus,
harmonic currents are diverted from their normal flow path on
the line through the filter.
Shunt passive filters depend on three design parameters: 1the tuning harmonic order, 2- the quality factor Q, the filter
band and the residual harmonic voltage depend on Q, 3- the
capacitor rating at the fundamental frequency.
Inductive reactance and resistor values at the fundamental
frequency are:

XL =

Xc
n2

R=

XL
Q

This is important point that the values of capacitor rating at

the fundamental frequency and the quality factor are discrete.
The typical value of quality factor is between 30 and 60 for

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
114

single-tuned notch filter. Moreover, the standard IEEE for

capacitor rating is available.
Based of new coding, in this work for each candidate buses
for installing passive filter, four filter branches are considered,
that each of these four filters is used for different harmonic
orders. The load buses that produce harmonics are considered
as candidate buses. For each of four filters seven bits are
considered in chromosome. The first bit shows presence or
absence, next three bits show the quality factor and the latest
three bits show the capacitor rating. With this coding method
the chromosome size is equal to 7*4*candidate buses.
III. OBJECTIVE FUNCTION
Objective function is the cost of filter branches. This cost
contains the cost of capacitors, inductors, and resistors. So:

fitfun = cprice + lprice + rprice

Where, cprice, lprice and rprice respectively are price of
capacitors, inductors and resistors. The cost of capacitors at
different voltage level in distribution networks is available. The
cost of filter inductors does not vary greatly for units of
different rating. The cost approximation used in the analysis is
of the form:

Inductor cost=UK+UL (MVAR)

Where UK is a constant cost component and UL is the
inductor incremental cost per MVAR rating. These values
relate to the single phase and must be converting to the three
phase.
The power rating of the resistor necessary for Q-adjustment
in each filter branch will affect the cost to some extent.
However, the nominal resistance of the unit is difficult to
predict in a general analysis, because it depends on the natural
Q factor of the inductor. For this reason, and because the cost
of an air-cooled resistor is small compared with that of the
other components, a constant cost per resistor is allocated in
the analysis [1].
For eliminating the individual that dont observe the
harmonic limitations, a penalty factor is used.
IV.

CONSTRAINS

Based on standard IEEE-519, the value of THD at PCC

(Point of Common Coupling) in distribution networks below
69 kV must be lower than 5% and the value of IHD lower than
3%. Also practically, the values of capacitor ratings and quality
factors are discrete.
V.

GENETIC ALGORITHM

The GA is a search mechanism based on the principle of

natural selection and population genetics. At the beginning of
GAs, representations for possible solutions, which are often
called chromosomes or individuals, must be developed. A
chromosome consists of a series of genes which can be
represented by binary codes. Different combinations of genes
form different chromosomes. Each chromosome is a possible
solution of the problem. The set of chromosomes is called the
population of the generation. Chromosomes in a generation are
forced to evolve toward better ones in the next generation by

three basic GA operators, reproduction, crossover, and

mutation, and the problem-specified fitness function.
In reproduction, a number of selected exact copies of
chromosomes in the current population become a part of the
offspring. In crossover, randomly selected subsections of two
individual chromosomes are swapped to produce the offspring.
In mutation, randomly selected genes in chromosomes are
altered by a probability equal to the specified mutation rate.
For a binary coding gene, it means that digit 1 becomes digit 0
and vice versa.
The fitness function, also called the objective function, is an
evaluation function that plays the role of the environment to
distinguish between good and bad chromosomes. Only a
certain number of chromosomes screened and selected by the
fitness function can survive and pass their genes to the next
generation.
Each chromosome of the current population is evaluated by
the fitness function and some good chromosomes are selected.
Then a new population will be generated by using the three
basic GA operators, i.e., reproduction, crossover, and mutation.
A. Reproduction
The reproduction of new individuals is made as follows:
- Find the total fitness of the population:
n

f total =

fi
i =1

Where:
fi : Fitness of each individual i(1,2,..., n )
n : Number of individual in the population
- Calculate the probability of reproduction for each
individual i (1,2,..., n ) .

pi =

fi
f total

- Round p i i (1,2,..., n ) and copy each individual i for

p i times at mating pool
- Enter three random individuals instead of last three
individuals of population
B. Crossover
The crossover operator follows:
- Select individuals two by two respectively
- Select a filter branch characters in random
Replace it by relative filter branch of next individual by the
probability of crossover ( pc )
C. Mutation
The mutation operator follows:
- Select individuals respectively
Reverse the bits relative to absence or presence of filter
branches by the probability of mutation ( pm )

VI.

CASE STUDY

This paper uses an IEEE standard distribution network to

perform the study. This network consists of 13 buses and is
representative of a medium-sized industrial plant. The plant is

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
115

fed from a utility supply at 69 KV and the local plant

distribution system operates at 13.8 KV [7].
The system is shown in figure 1. More information on this
network
can
be
obtained
at
[Link]

50:GEN-1

1
7
26:FDR G
150
50
2.1
68.8
3.63

Table 3
THD of buses before and after filtering

01:69-1

03:MILL-1

26:FDR G

06:FDR H

29:T11 SEC

49:RECT

Filter branch
Tuning order
Location
XC (KVAR)
Q
C (F)
L (mH)
R( )

100:UTIL-69

51:AUX

05:FDR F

Table 2
Simulation result

39:T3 SEC

11:T4 SEC

19:T7 SEC

ASD

Figure 1: IEEE 13-buses standard network

bus
50:GEN-1
03:MILL-1
51:AUX
100:UTIL-69
01:69-1
05:FDR F
49:RECT
39:T3 SEC
26:FDR G
29:T11 SEC
06:FDR H
11:T4 SEC
19:T7 SEC

Initial THD (%)

10.56
10.93
9.95
1.34
1.36
10.93
11.04
9.86
10.92
11.12
10.93
10.55
9.96

Final THD (%)

2.48
2.55
7.72
0.32
0.33
2.55
10.23
2.32
2.56
10.41
2.55
2.47
2.34

In this paper in addition to 49:RECT, the 29:T11SEC and

51:AUX is considered also as a nonlinear load.
Standard rating of single-phase capacitors and their costs
that released by ABB (according to IEEE and IEC standard),
are shown in following table:
Table 1
Standard single-phase capacitor ratings and their costs
XC (KVAR)
50
100
150
200
300
400
500
600

PRICE ($)
550
625
705
790
950
1155
1315
1455

In this paper the value of UK and UL is considered 500$

and 90 ($/MVAR) respectively. p m , pc and the size of
population are considered 0.15, 0.75 and 10 respectively.
VII. SIMULATION RESULT
Passive filters locate at the primary side of load transformers
(PCCs) generally, thus in this problem, the candidate buses for
filters are 50:GEN-1, 05:FDR F and 26:FDR G.
The program result is shown in table 2 and Table 3 shows
the THD values of buses before and after filtering.
Convergence curves of the algorithm for five individual
runs are shown in figure 2.

Figure 2: Convergence curves for five algorithm runs

Curve A and B in fig.2 show the changing of transfer

impedance between 26:FDR and 29:T11 SEC buses
respectively before and after installing passive filter. Because
of using of power factor correcting capacitor in network, as
shown in fig.3 network has a resonance frequency near the 7th
harmonic order. After installing filter this frequency change to
two new resonance frequencies far from 7th harmonic order.
In designing passive filters for a network it must be
considered that after installing filters new resonance frequency
near the nonlinear loads order dont be appeared. However,
with using genetic algorithm by consideration a penalty factor
in fitness function appearing of this new resonance frequency
is prevented.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
116

that define by standard IEEE-519. The algorithm determines

the number, harmonic order, size and location of passive filters
to reach the IEEE standard..
REFERENCES
[1]
[2]
[3]
[4]

[5]
Figure 3: Harmonic impedance before (A) and after (B) filtering
[6]

VIII.

CUNCLOSION

This paper deals with determining the location and size of

passive filters by genetic algorithm, economically. The
algorithm constrains are the limitation of harmonic distortion

[7]

J. Arrillaga, N.R. Watson, Power System Harmonics, Second Edition,

2003, John Wiley & Sons
J.S. Subjak and J.S. Mcquilkin, Harmonics- Causes, Effects,
Measurements, and Analysis: An Update, IEEE Trans. Ind. Applicant. ,
Vol.26, pp. 1034-1042, 1990.
Yaow-ming Chen, Passive Filter Design Using Genetic Algorithm,
IEEE Transaction On Industrial Electronics, Vol. 50, No. 1, Feb 2003.
C.C.M. Moura, M.E.L. Tostes, E.P. Santos, R.C.L. Oliveria, T.M.M.
Branco, U.H. Bezerra, Determination of the R-L-C Parameters of a
Passive Harmonic Filter Using Genetic Algorithm, IEEE Conf. on
Harmonics and Quality of Power, Vol. 2, PP. 495- 500, 2002.
Franklin M.P. Pamplona, Benemar A. Souza, Harmonic Passive Filter
Planning In Radial Distribution Systems Using Microgenetic Algorithm,
IEEE 11th International Conf. on Harmonics and Quality of Power, 2004.
A. Berizzi, C. Bovo, The Use of Genetic Algorithms for the Localization
and the Sizing of Passive Filters, 9th International Conf. on Harmonics
and Quality of Power, Vol. 1, PP. 19-25, 2000.
Task Force on Harmonics Modeling and Simulation, Test Systems for
Harmonics Modeling and Simulation, IEEE Trans. On power Delivery,
Vol. 14, No. 2, April 1992.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
117

Balancing Voltage Source Direct Torque Control of

Induction Motor with Three level Inverter
Hacene Rezine , Abderrahmane Berkani .

UER Electrotechnique, EMP (Ex-ENITA). BP 17 Bordj-El-Bahri, Algiers, Algeria

Email : Abdouberkani@[Link]
Rezine_hacene_emp@[Link]
Abstract In this paper we present a new control structure for
sensorless induction machine dedicated to electrical drives using a
three-level voltage source inverter (VSI). The amplitude and the
rotating speed of the flux vector can be controlled freely. Both fast
torque response and optimal switching logic can be achieved; the
selection is based on the value of the stator flux and the torque.
A novel DTC scheme of induction motors is proposed in order
to develop a suitable dynamic. We propose an approach; in wich
we enhance the response of torque and flux with optimal
switching strategies. However, the middle point voltage of the
input DC voltages of the three-level NPC voltage source inverter
presents serious problems caused by a fluctuation of the DC
voltage sources Ucu Ucl As consequence to these problems, we
obtain an output voltage of the inverter which is asymmetric and
with an average value different from zero. In this paper, we will
present one solution to minimise this fluctuation. This solution
uses a clamping bridge to regulate the input voltages of a
threelevel inverter VSI NPC. A scheme of Enhanced direct torque
control "EDTC" with complete cascade is simulated for an
induction motor. The results obtained indicate superior
performance over the FOC one without need to any mechanical
sensor.

the stator flux. The direct torque control (DTC) is one of the
actively researched control scheme which is based on the
decoupled control of flux and torque providing a very quick
and robust response with a simple control construction in ac
drives [3].
In this paper, the authors propose a new cascade for high
voltage and high power applications. This cascade lets to
absorb, in network, sinusoidal currents with unity power factor.
It constitutes by two-level PWM rectifier, clamping bridge,
three-level NPC VSI, induction motor controlled by EDTC
strategy.
II. THREE-LEVEL INVERTER TOPOLOGY
Fig. 1 shows the schematic diagram of neutral point clamped
(NPC) three-level VSI. Each phase of this inverter consists of
two clamping diodes, four GTO thyristors and four
freewheeling diodes. Table.I shows the switching states of this
inverter. Since three kinds of switching states exist in each
phase, a three level inverter has 27 switching states.

Key Words - Induction motor, Direct torque control, Voltage

source inverter, , Flux estimators, Switching strategy optimisation,
Neutral-point clamped.

I.

INTRODUCTION

The rapid development of the capacity and switching

frequency of the power semiconductor devices and the
continuous advance of the power electronics technology have
made many changes in static power converter systems and
industrial motor drive areas. The conventional GTO inverters
have limitation of their dc-link voltage. Hence, the series
connections of the existing GTO thyristors have been essential
in realizing high voltage and large capacity inverter
configurations with the dc-link voltage [1]. The vector control
of induction motor drive has made it possible to be used in
applications requiring fast torque control such as traction [2].
In a perfect field oriented control, the decoupling
characteristics of the flux and torque are affected highly by the
parameter variation in the machine.
This paper describes a control scheme for direct torque and
flux control of induction machines fed by a three-level voltage
source inverter using a switching table. In this method, the
output voltage is selected and applied sequentially to the
machine through a look-up table so that the flux is kept
constant and the torque is controlled by the rotating speed of

Fig. 1. Schematic diagram of a three-level GTO inverter

Table. I.
Switching states of a three-level inverter
S1

S2

S3

S4

ON

ON

OFFF

OFFF

Vd

OFFF

ON

ON

OFFF

Vd/2

OFFF

OFFF

ON

ON

Switching
states

SN

A 2 -level inverter is only able to produce six non-zero voltage

vectors and two zero vectors [2]. The representation of the

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
118

space voltage vectors of a three-level inverter for all switching

states forming a two-layer hexagon centred at the origin of the
(d, q) plane and a zero voltage vector at the origin of the plane,
as depicted in fig.2. According to the magnitude of the voltage
vectors, we divide them into four groups the zero voltage
vectors (V0), the small voltage vectors (VI, V4,V7, VI , V13 ,
V16 ), the middle voltage vectors (V3 , V6 V9, V12 V15 V18 ),
the large voltage vectors (V2 , V5 , V8 , V11,V14, V17 )
The zero voltage vector (ZVV) has three switching states, the
small voltage vector (SVV) has two and both the middle
voltage vector (MVV) and the large voltage vector (LVV) have
only one [1].

Ph2
V8 =(020)

V6 =(120)

V4=(110)

V7 =(121)

V4 =(221)

V0 =( (000)

V2=(200)

V12= (012)

V0 = (111)
V0 = (222)

Ph1

V1=(211)

V13 =(001)

V16 =(101)

V13 =(112)

V16 =(212)

V14 =(002)

Basically, DTC schemes require the estimation of the stator

flux and torque. The stator flux evaluation can be carried out
by different techniques depending on whether the rotor angular
speed or (position) is measured or not. For sensorless
application, the "voltage model" is usually employed [6]. The
stator flux can be evaluated by integrating from the stator
voltage equation.
(6)
s (t ) = (V s R s I s )dt

V1=(100)

V11 =(022)
V10 =(122)

IV. PROPOSED DIRECT TORQUE CONTROL USING A

THREE-LEVEL INVERTER

V3=(210)

V9 =(021)

V10 =(011)

electromagnetic torque developed by the machine.

V5=(220)

V7 =(010)

d
(5)
= em r ()
dt
r () and em are respectively the load torque and the
J

V18 =(201)

V15= (102)

V17 = (202)

Ph3

Fig. 3. Flux deviation

Fig. 2. Space voltage vectors of a three-level inverter

III. INDUCTION MACHINE

Torque control of an induction motor can be achieved on the
basis of its model developed in a two axis (d, q) reference
frame stationary with the stator winding. In this reference
frame and with conventional notations, the electrical mode is
described by the following equations:
di sd
1
p
1 1 1
=
sd +
sq + i sd p i sq
dt
T r Ls
Ls
T r Ts
1
+
V
Ls sd

di sq
dt
+

T r Ls

sq +

Ls

1 1

V cos(v s )
s T e s
(1)

sd + i sq + p i sd
T r Ts
(2

1
V
Ls sq

This method is very simple requiring the knowledge of the

stator resistance only. The effect of an error in Rs is usually
neglected at high excitation frequency but becomes more
serious as the frequency approaches zero.
The deviation obtained at the end of the switching period Te
can be approximate by the first order Taylor Series as below
[6].
(7)
V T cos( )

(8)

so

Considering the combination of states of switching functions

Su, Sv, Sw. Fig.3 shows the adequate voltage vector selection,
we can increase or decrease the stator flux amplitude and phase
to obtain the required performance. The electromagnetic torque
is estimated from the flux and current information as [2]:

em = p (i sq sd i sd sq )

d sd
= V sd R s i sd
dt
d sq
= V sq R s i sq
dt

(3)
(4)

The mechanical mode associated to the rotor motion is

described by:

Fig.4. shows a block diagram of the DTC scheme developed by

I. Takahashi [2]. The reference values of flux,

torque

*
em

*s

, and

, are compared to their actual values and the

resultant errors are fed into a multi-level comparator of flux

and torque. The stator flux angle s is calculated by:

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
119

s = arctan

sq
sd

(9)

and quantified into 6 levels depending on which sector the flux

vector falls into. Different switching strategies can be
employed to control the torque according to whether the flux
has to be reduced or increased.
Each strategy affects the drive behavior in terms of torque and
current ripple, switching frequency and two or four quadrant
operation capability.
Assuming the voltage drop ( R s

I s ) small, the head of the

stator flux s moves in the direction of stator voltage Vs at a
speed proportional to the magnitude of Vs according to:
(10)
s =V sT e
The switching configuration is made step by step, in order to
maintain the stator flux and torque within limits of two
hysteresis bands. Where Te is the period in which the voltage
vector is applied to stator winding. Selecting step by step the
voltage vector appropriately, it is then possible to drive s
along a prefixed track curve [6].

vectors having large components tangential to the stator vector

locus, implies very high switching frequency.
For flux control, let the variable

E (E = s* s ) be

located in one of the three regions fixed by the contraints:

E p E min , E min E E max , E f E max

(11)

The suitable flux level is then bounded by E min

and

E max Flux control is made by a two-level hysteresis

comparator.
Three regions for flux location are noted, flux as in fuzzy
control schemes, by E n (negative), E z (zero) and E p
(positive).
A high level performance torque control is required. To
improve
the
torque
control,
let
the
difference

(E = *em e ) belong to one of the five regions defined

by the contraints:

E p Emin2 , Emin2 E Emin1 , Emin1 E Emax1 ,

(12)

Emax1 E Emax2 and Emax2 p E

The five regions defined for torque location are also noted, as
in fuzzy control schemes by Enl (negative large), Ens
(negative small),

Fig. 4. Block diagram of direct torque control

Assuming the stator flux vector lying in the k-th sector

(k=1,2,...6) of the (d, q) plane, in the case of three-level
inverter, to improve the dynamic performance of DTC at low
speed and to allow four-quadrant operation, it is necessary to
involve the voltage vectors Vk 1 and Vk 2 n torque and flux
control.

(positive large). The torque is then controlled by a hysteresis

comparator built with two lower bounds and two upper bounds.
A switching table is used to select the best output voltage
depending on the position of the stator flux and desired action
on the torque and stator flux. The flux position in the (d, q)
plane is quantified in twelve sectors. Alternative tables exist
for specific operation mode. Comparing with switching table
for the case of a two-level inverter, it is easily possible to
expand the optimal vector selection to include the larger
number of voltage vectors produced by three-level inverter
[4] ,[5]. The appropriate vector voltage is selected in the order
to reduce the number of commutation and the level of steady
state ripple.
The switching strategy in the order of the sector s , is
illustrated by each table. The flux and torque control by vector
voltage has in nature a desecrate behavior. In fact, we can
easily verify that the same vector could be adequate for a set of
value of s .

V. SWITCHING STRATEGY PROPOSED FOR AN

ENHANCED DIRECT TORQUE CONTROL
According to this strategy, the stator flux vector is required to
rotate in both positive and negative directions. By this, even at
very low shaft speed, large negative values of rotor angular
frequency can be achieved, which are required when the torque
is to be decreased very fast. Furthermore, the selection strategy
represented in each table allows good flux control to be
obtained even in the low speed range. However, the high
dynamic performance, which can be obtained using voltage

Ez (zero), Eps (positve small), Epl

cflx
ccpl
2
1
-1
-2

ce
1
-1
1
-1
1
-1
1
-1

2
V5
V5
V4
V4
V7
V7
V8
V8

1
V4
V4
V4
V4
V7
V7
V7
V7

-1
V16
V16
V16
V16
V13
V13
V13
V13

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
120

-2
V17
V17
V16
V16
V13
V13
V14
V14

2
VI. CAPACITOR VOLTAGE BALANCING ALGORITHM

cflx
ccpl

ce
1
-1
1
-1
1
-1
1
-1

2
1
-1
-2

2
V8
V8
V7
V7
V10
V10
V11
V11

1
V7
V7
V7
V7
V10
V10
V10
V10

-1
V1
V1
V1
V1
V16
V16
V16
V16

-2
V2
V2
V1
V1
V16
V16
V17
V17

3
cflx
ccpl

ce
1
-1
1
-1
1
-1
1
-1

2
1
-1
-2

2
V11
V11
V10
V10
V13
V13
V14
V14

1
V10
V10
V10
V10
V13
V13
V13
V13

-1
V4
V4
V4
V4
V1
V1
V1
V1

-2
V5
V5
V4
V4
V1
V1
V2
V2

ce
1
-1
1
-1
1
-1
1
-1

2
1
-1
-2

2
V14
V14
V13
V13
V16
V16
V17
V17

1
V13
V13
V13
V13
V16
V16
V16
V16

-1
V7
V7
V7
V7
V4
V4
V4
V4

-2
V8
V8
V7
V7
V4
V4
V11
V11

-1
V10
V10
V10
V10
V7
V7
V7
V7

-2
V11
V11
V10
V10
V7
V7
V8
V8

5
cflx
ccpl

ce
1
-1
1
-1
1
-1
1
-1

2
1
-1
-2

2
V17
V17
V16
V16
V1
V1
V2
V2

1
V16
V16
V16
V16
V1
V1
V1
V1

of the first sub-group, this vectors obtained with combinations

where 1 or 2 states is equal 1, the remaining states being 0,
causes to increase the neutral point voltage Uo of the inverter.
On the other hand, the use of the vectors of the other sub-group
of the "SVV" group, vectors obtained for combinations where
1 or 2 states is 1, the remaining states being 0, causes to
decrease it (opposite effect).
Thus, if one decided to use only the vectors of one of
the two sub-groups, in this case the direct control then become
equivalent to the direct control of the couple of the induction
machine supplied with a 2-levels inverter where DC voltage is
equal to the one of the voltages of the condenser.
However the input voltage of the 3-levels inverter is
maintained constant (13) and is equal:

V d = U Cu + U Cl

4
cflx
ccpl

We know that the group small voltage vector "SVV"

is divided into two sub-groups of vectors which do not have the
same action on the voltage of neutral point voltage of the
inverter. On the assumption that U 0 = U Cl , use of the vectors

(13)
The use of the vectors of the first sub-group will have
for effect to increase the value of the neutral point voltage Uo .
However if we chose U 0 = U Cl , the terminal voltage of the
condenser Cl will increase and that of the condenser Cu will
decrease until being cancelled. It is thus not possible to control
the machine by using only the vectors of one sub-group
If we take for example the two vectors V4 and V4'
which belong each one to one of the two sub-groups of the
" SVV " group. They have both the same action on the module
of stator flux and the value of the electromagnetic torque. On
the other hand, they have contrary actions on the value of the
neutral point voltage Uo. The first will increase it while the
second will decrease it. With this property one can control the
variations of Uo by maintaining them in a band of hysteresis
centered around the value of reference which is null.
After having measured the voltages of the two
condensers, we can calculate the variation (14) which exists
between the two and which corresponds to an imbalance of the
two condensers:
(14)
= U U
t

cflx
ccpl
2
1
-1
-2

ce
1
-1
1
-1
1
-1
1
-1

2
V2
V2
V1
V1
V4
V4
V5
V5

1
V1
V1
V1
V1
V4
6V4
V4
V4

-1
V13
V13
V13
V13
V10
V10
V10
V10

-2
V14
V14
V13
V13
V10
V10
V11
V11

Cl

Cu

We use hysteresis controller with 2 states to maintain the error

in a band of hysteresis of width. 2 . V d
If,

t < V d or

if

t > V d

that means that

imbalance between the two condensers becomes unacceptable

in a direction or the other. In the first case, it will be necessary
to impose a voltage vector which will have as an action to
increase the neutral point voltage Uo of the inverter; the
second, one will impose a voltage vector which will have as an
action to decrease it.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
121

20
Courant (A)

Couple (Nm)

40
20
0

-20
0

0.05

0.1

0.15

0.18

0.19

0.2

0.21

1000
Tension (V)

1.5
Flux (Wb)

-20
0.17

0.2

1
0.5
0
0

0.05

0.1
Temps (s)

0.15

0.2

-1000
0.3

0.305

0.31

0.315 0.32
Temps (s)

0.325

0.33

Fig. 7. Performances of the DTC drive associated to its supply at nominal

reference of the flux and torque.

Fig. 5. Voltage unbalancing of the inverter capacitors

VII. THE SIMULATION RESULTS

The goodness of the proposed DTC algorithm has been verified
in simulation on a PC. All simulations have been performed in
the Matlab/Simulink environment. The used flux and torque
contraints are expressed in percent with respect to the flux and
torque reference values.

E max = 3% , E min = 3% , E min1 = 0.8%

E min 2 = 3% , E max1 = 0.8% , E max 2 = 3%

VIII.

The simulation result illustrates both the steady state

and the transient performance of the proposed torque control
scheme.
A

480

10
Id1 (A)

Vc1 (V)

460
440
420

0.1

0.2

0.3

0.4

0.5

-20
0.15

0.6

480

20

460

10
Id2 (A)

Vc2 (V)

0
-10

400

440
420

0.155

0.16

0.165

0.17

0.175

0.18

0.155

0.16

0.165

0.17

0.175

0.18

0
-10

400
0

0.1

0.2

0.3

0.4

0.5

-20
0.15

0.6

100

50

50
I0 (A)

Vc1-Vc2 (V)

20

-50
0

0.1

0.2

0.3
0.4
Temps (s)

0.5

0.6

CONCLUSION

This paper presents a very simple method of the Direct Torque

Control algorithm to be applied to 3-level Diode clamped
inverters. The effect of proposed method has been proved by
simulations. It is concluded that the proposed control produces
better results for transient state operation. And it is suitable for
high-power and high-voltage applications. We enhance the
DTC approach by introducing two multi-level hysteresis
comparators for flux and torque control. We impose the flux
angle detection procedure by defining twelve sectors of space
and establish a larger table of knowledge rules with optimal
switching strategies.
Also, the neutral point voltage can be easily controlled by
introducing a clamping bridge to solve the usual problem of
unbalanced voltages input in three-level [Link]. From this
analysis high dynamic performance, good stability and
precision are achieved; the results obtained are full of promise
to use this system in high voltage and great power applications
as electrical traction.
IX. REFERENCES

-50
0.15

from figure (6-A), It can be observed that with the

control algorithm the voltages in the capacitors C1 and C2
follow their references and the and difference between them is
practically low, On the figure (6-B) we can see that the inputs
currents in the inverter Id1 et Id2 have the same form but they
have are in phase opposition, the mean value of the Id0 current
are equal to zero.
From The figure (7) ,the flux the torque oscillates around their
reference values (respectively 1 Wb and 50 N.m), a few torque
ripples, and the response time is very short taking into account
the time add by the use of the algorithm without any need of
more powerful [Link] can see that both of The voltage
and current qualities (wave forme , harmonics) are very good
which prove also the utility of this simple algorithm .

0.155

0.16

0.165 0.17
Temps (s)

Fig. 6. A) DC input voltage of the inverter Vc1-Vc2

and their difference Vc1-Vc2
B) input currents of the 3-level inverter

0.175

0.18

[1] Y.H. Lee, B.S. Suh, and D.S. Hyan, A novel PWM
scheme for a three-level voltage source inverter with GTO
thyristors. IEEE Trans. on Ind. Appl, vol. 33 2, March\April
1996, pp. 260-268.
[2] I. Takahashi and T. Noguchi, A new quick-response and
high-efficiency control strategy of an induction motor. IEEE
Trans. on IA, vol. 22, No. 5, Sept\Octo 1986, pp. 820-827

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
122

[3] J.C. Trounce, S.D. Round, and R..M. Duke, Comparison

by simulation of three-level induction motor torque control
schemes for electrical vehicle applications. Proc. of
international power engineering conference,2007 IEEE .
[4] R. Zaimeddine, and E.M. Berkouk, A Novel DTC scheme
for a three-level voltage source inverter with GTO thyristors.
SPEEDAM 2004, Symposium on power electronics, electrical
drives, automation & Motion, June, vol. 2, June, 16th-18th
2004, pp. F1A-9-F1A-12.
[5] R. Zaimeddine, and E.M. Berkouk, A Scheme of EDTC
Control using a Three-Level Voltage Source Inverter for an
Induction Motor. Proc. of international power engineering
conference IEEE 2007
[6] F. Bacha, A. sbai, and R. Dhifaui, Tow Approaches For
Direct Torque Control of an Induction Motor. CESA
Symposium on control, vol. 1, March 1998, pp. 562-568.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
123

REMOTEI
RRI
GATI
ONCONTROLLER
UNDERPHOTOVOLTAI
CENERGY
M. L. BENZAOUI*, F. YOUCEF ETTOUMI, A. ADANE

University of Science and Technology Houari. BOUMEDIENE (USTHB)

Faculty of Electronic & Computing
BP 32, El-Alia, Bab-Ezzouar 16111, Algiers, Algeria
Fax : 213 21 21 82 14
*Email : benzaoui_m_l@[Link]
Abs
t
r
ac
t
-Wa
t
e
ri
sani
mpor
t
antt
hi
ngi
nt
hee
a
r
t
h,s
owemus
tus
e
i
te
c
o
nomi
c
a
l
l
y,e
s
pe
c
i
a
l
l
yi
nt
hei
r
r
i
ga
t
i
ono
ff
i
e
l
dsi
nt
hedr
y
a
r
e
a
sa
ndi
ns
umme
rda
ys
.Sot
ha
twepr
e
s
e
nti
nt
hi
spape
r
,a
c
onc
e
pt
i
ono
fade
v
i
c
ewhi
c
hpe
r
mi
t
sust
oc
ont
r
o
laut
oma
t
i
c
a
l
l
y
t
heus
e
so
fwa
t
e
ri
ni
r
r
i
ga
t
i
o
[Link]
si
sbymi
c
r
o
c
ont
r
o
l
l
e
r
.We
a
ppl
i
e
dt
hi
sde
v
i
c
ei
nt
heSa
ha
r
ao
fAl
ge
r
i
a(
oa
s
i
s
)
.
Andt
oge
taut
onomyo
ft
hes
ys
t
e
m,weha
v
epr
e
f
e
r
r
e
dt
ous
et
he
Pho
t
o
vo
l
t
a
i
cs
ys
t
e
msa
sane
l
e
c
t
r
i
c
a
ls
uppl
yo
ft
hei
ns
t
a
l
l
a
t
i
o
nand
apo
we
rf
o
rwa
t
e
rpumpi
ng.
Thegr
e
e
tpr
obl
e
mi
ni
r
r
i
ga
t
i
oni
st
os
pe
c
i
f
ywa
t
e
rne
e
dsf
o
re
a
c
h
pl
ant
,whi
c
hde
pe
ndso
ft
e
mpe
r
a
t
ur
e
,l
a
ndhumi
di
t
y,a
i
rhumi
di
t
y,
wi
nda
ndt
hena
t
ur
eo
ft
hepl
ant
.
I
nf
i
r
s
tweha
v
eus
e
di
nt
hec
ont
r
o
l
l
e
rde
v
i
c
emanys
e
ns
or
st
oge
t
t
hei
nf
o
r
mat
i
onsaboutt
he
s
epa
r
ame
t
e
r
swhi
c
ha
r
eus
e
da
si
nput
s
o
ft
het
e
s
tc
ondi
t
i
o
nst
oi
r
r
i
ga
t
e
.
And ne
xta
ne
l
e
c
t
r
oni
c de
v
i
c
e ba
s
e
d on t
he mi
c
r
o
c
o
nt
r
ol
l
e
r
16F87
7wa
sr
e
a
l
i
s
e
di
no
r
de
rt
oge
tan aut
oma
t
i
cc
ont
r
o
la
nd
c
ommands
.
Fi
ne
l
yweha
vet
e
s
t
e
dt
hi
sde
v
i
c
et
oi
r
r
i
ga
t
et
omat
o
,be
t
we
e
nJune
a
ndJul
y,a
ndr
e
s
ul
t
swe
r
ego
o
d.
Ke
yWo
r
ds
:I
r
r
i
ga
t
i
on,
wa
t
e
rpumpi
ng,Pho
t
o
vo
l
t
a
i
c
,
mi
c
r
oc
ont
r
o
l
l
e
r
.

I.

Underground irrigation.
Irrigation under pivot.
Irrigation by sprinkling.
Irrigation drop by drop.

In our work we have chosen the irrigation drop by drop,

because its economic and easy.
For that we need some information of the local climate
(microclimate), the plant type and the soil nature [3],[4]. We
stocked water in a reservoir by using an electro-valve and
pump on the water supply.
An electronic devise based on the microcontroller 16F877 is
used with some sensors to survey the water level at the
reservoir and activate irrigation (figure1).
And to get autonomy of the system, we used a photovoltaic
generator as a power supply.
Sensors

Microcontroller

Irrigation

INTRODUCTION

In rural and remote area, to overcome the lack of water, the

photovoltaic (PV) water pumping represents one of the best
solutions. This is because solar radiation is generally high in
arid zone location where the water is most needed [1].
Also, uses of water are a big problem in these days. The
great consumption is in the Industrial uses and the irrigation.
So we must find methods to reduce its exploitation.
In general when we irrigate plants we try to see the sol if its
dry, humid and if the climate is hot. We usually irrigate using a
greet quantity of water and without taking in consideration if
the plants needs really water.
In this paper we propose to use a remote device to controller
the water pumping and irrigation with photovoltaic supply.
II. IRRIGATION METHODS
There are many methods used in irrigation, as example we can
find [1], [2]:

Irrigation of surface.

Photovoltaic
generator

Figure 1: Irrigation System.

III. SENSORS
Information about the microclimate such as the temperature,
Soil Humidity, air Humidity are important to get the best way
of water uses, also we must know the water level in the
reservoir. We have used for this [5]:

[Link]
mpe
r
at
ur
e
To get the air temperature and the soil one, we have used two
CTN resistors as sensors. These measures are compared with a
reference temperatures Tref fixed by the user (figure 2). The
comparator is a low power dual operational amplifier (LM358).
The outputs are connected to the controller

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
124

[Link]
c
r
oc
ont
r
ol
l
e
r

+5V
CTN

LM358

+5V

Output

All these measures are used as inputs of a microcontroller, to

get the right time for the irrigation and fill the reservoir. This
microcontroller is the PIC 16F877 of MICROCHIP [6]. It has
the advantage that its easy to use and cheaper and sufficient to
our application.

Tref

IV IRRIGATION
Figure 2: Temperature measurement
[Link]
di
t
y
In order to use water when its necessary (for example the soil
is dry), we have to know the humidity of the air and the
humidity of the soil, so we have used tow sensors
(232269190001 of Philips), which are considered as a variable
capacitor depended of the humidity.
So we placed these sensors in a Timer (NE555), to generate
a cycle of time depended on humidity (figure 3). We use the
microcontroller to activate the Timer and calculate the
humidity by the measure of =f(CH).

NE555
CH

Output
=f(CH)

For irrigation we have to use a pump (P1) and a solenoid-valve

(EV1) activated by the microcontroller to fill the reservoir. A
second one (EV2) used to evacuate water from the reservoir
(irrigation) and an other pump (P2) used in irrigation when the
reservoir is empty.
We added a state where the reservoir is nearly empty in order
to reduce the quantity of water used in irrigation to get more
time when the water is enough.
So we have used 4 water levels:
Full 100%: turn Off EV1 and pump1 and turn On EV2,
the irrigation is normal.
Half empty (50%): turn On P1, EV1 and we have to turn
On EV2 in an economical irrigation.
Nearly empty (20%): turn On P2, EV1 and EV2 used in
the necessitate irrigation.
Empty (0%): turn Off the pump P2 and EV2, turn On EV1
we can not irrigate.
The pumps used are a PP/MH-809-12V (5.5 gal/mn).
And the solenoid-valve are EV1L280 [7].
All the irrigation installation is presented in figure 6.

Figure 3: Humidity measurement

Pump P1
Water
Supply

[Link]
e
rl
e
v
e
l
When we use the water from the reservoir we can fill or empty
it, so we must know the level of the water in the reservoir. For
this we have used 4 opto-transistors with 4 LED. In order to
measure the 4 water levels.

N4

N3

+5V

+5V

N2

LED
Wate
r

EV1

Tank

EV1 Pump

N1
Ta

Output

Irrigation
Ha

a) Water resources

b) Irrigation drop by drop

Hs
Ts
%
Parameter

Reservoir
Water

Figure 4: Water level measurement

Figure 6: Irrigation system.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
125

V PHOTOVOLTAIC SYSTEM
All the components of our system are chosen to be used under
12V DC, this is in order to use battery and a Photovoltaic (PV)
installation.
The photovoltaic generator used is a 12V/50W (PWX500-50 of
PHOTOWATT), and the battery is 12V/65AH (VARTA
SOLAR) [8].

Finely we have tested this installation to irrigate tomato, in

summer (between June and July), at Dj
amaaa town in the
Sahara of Algeria, where there is luck of water, the weather is
very hot and dry. We have not used the green houses in order
to get the real climate.
The results were good and the plants of tomato has resisted to
the climate.

So the power supply is as in figure 7.

K1
Photovoltaic
Generator

VI CONCLUSION
Regulator

K3

+
K2
-

Load
Battery

Figure 7: the photovoltaic power supply systeme.

AREGULATOR
In this installation Fig 2, the major problem is in the uses of
battery, because it can be damaged easily by the charging
currant, the cycle of charges number, and if we charge it up to
its limit (charge exceed) or to discharge it under its limit
(profound discharge). These thresholds are fixed by the battery
characteristics [1]. To resolve these problems, we use a
regulator which can be considered as controlled switches K1,
K2 and K3 [9]. These have to keep the battery by
disconnecting it from the photovoltaic generator if the battery
potential is up to its limit, or to disconnect battery from uses if
its potential is under its limit [1],[9] and to prolong the battery
charge and discharge state and eliminate the quick
commutations of switches [9].

The aim problem at the Sahara and the Arab countries is water
resource, so we have to use it economically especially in the
industry and in irrigation. We always, use pump in these
regions, which need power supply, so photovoltaic become one
of the best solutions to electrify arid areas. We can say that our
application is very important to generate it to conserve our
reserve of water.
References
[1] Med. L. BENZAOUI, F. YOUCEF ETTOUMI and A. ADANE,
Modelling of Photovoltaic Regulator with MPPT. Proceedings of the
International Conference on Modeling and Simulation (MS07 Algeria) July 2 4, 2007, Algiers, Algeria
[2] CEMAGREF, Guide pratique d'irrigation, Dunod 1992
[3] Zairi A., Mailhol J. C., Nasr Z., Slatni A., & Ruelle P. 1999. Surface
irrigation strategies and their impacts on yield and water consumption - The
Tunisian case. 50th International Executive Council Meeting, 17th
International Congress of ICID. Q 48/1.31, September 11-19, 1999, Granada,
Spain. pp. 42-50
[4] I. Broner, Drip Irrigation for Home Gardens Gardening series no. 4.702.
[5] G. ASCH Les capteurs d'instrumentation industrielle. Dunod 1989
[6] C. Travenier Application industrielles des PIC. Edition Dunod 2000
[7] International Product Catalogue, A Siebe Group Company.
[8] Natural Resources Canada, Photovoltaic Systems: A Buyers Guide,
catalogue N : M92-28/2001F ISBN : 0-662-86306-2.
[9] Med. L. BENZAOUI, F. YOUCEF ETTOUMI and A. ADANE,
Simulation of water pumping using photovoltaic energy. Proceedings of the
International Conference on Electrical Engineering Design and Technologies
ICEEDT08 HAMMAMET, TUNISIA November 8-10, 2008
.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
126

Implementation of Inter-area Angle Stability

Prediction in Wide Area Control
Ahmed A. Daoud
Abstarct: This paper proposes a new interarea angle
stability prediction algorithm. This proposed algorithm
does not require any prior knowledge of system state as it
operates directly from measurements drawn from PMUs.
The proposed predictor foresee the system stability state
for 500 ms in advanve. Applying the COI angle concept for
the interconnected power systems, inter-area stability can
be predicted in a proper time. The proposed method was
applied for two standard test systems, two area 4
generator system and IEEE 50 generators test system. The
results showed that the prediction accuracy in most cases
converges around 1 %. At the worest case the predicted
values approaches 7%.

1. INTRODUCTION

ower systems are large interconnected nonlinear

systems. They respond to a disturbance over a
varying timescale, ranging from milliseconds even to
hours. Wide range of contingencies and disturbances
occur and abnormal operating conditions may be
detected. A different corrective control schemes must be
implemented in each type of contingency or disturbance
at the right time [1].
Because of deregulation, many power systems around
the world are being forced to operate closer to their
stability limit because of the operational requirements in
an open access environment and the environmental
considerations. Power transfers across the interconnected areas of a power system are unpredictable
due to market price variations. Inter-connection
unforeseen operating conditions and area instability can
lead to system blackout [2].
The recent series of blackouts in different countries has
further emphasized the need for operators to have better
information regarding the actual state of the power
system they are operating.
For the last two decades, advances in computer
technology and communications provide the operators
with the information needed for appropriate control
action. At the same time, measurement systems based
on phasor measurement units (PMUs) are becoming
proven technology and are seen by many utilities as one
of the most promising ways of providing phasor
information for wide area control [3].
_____________________________________________
Dr. Ahmed A. Daoud is with the electrical engineering
dept., Suez Canal University, Port Fouad, Port Said,
Egypt, 42523
E-mail: [Link]@[Link]

Since many real-time operating decisions, both manual

and automatic, are based on software applications using
information derived from the phasor measurements
through communication, these developments have
shown immediate benefits in terms of increased
accuracy, stability, and speed of convergence of control
action decision making.
Current software and algorithms used in wide-area
control are based on phasor measurements of bus
voltage and generator reactive power [1]. In some
applications, it is effective to use phase angle
measurements to detect inter-area angle instability [1].
2. PROBLEM FORMULATION
If a severe disturbance, such as three-phase
fault occurs in a tie-line between two areas of an
interconnected power system, some of the generators of
the power system may accelerate and may lose
synchronism. If such a disturbance is not cleared at a
proper time, the loss of synchronism may extend to
other areas' generators and a blackout may occur [4].
The loss of synchronism after fault clearing is affected
by the increase or decrease of the generator relative
rotor angle beyond an identifiable threshold.
Phase angle prediction can be used to detect the first
swing instability in advance, which gives time to
operator to apply the proper preventive control action.
Based on phase angle prediction and frequency
measurements of critical generators buses from the
entire interconnected power system transient stability
can be detected and mitigated.
3. PROPOSED ALGORITHM
The phase angle of generator bus and relative
phase difference varies through a wide range during
system operation. The concept of center of inertia (COI)
for the computation of the system phase angle reference
is used to determine the interconnection phase angle [3].
This approach is used to quantify the extent of phase
angle variations away from the system center.
Since the internal rotor angle cannot directly measured,
we approximate the internal angle with the phase angle
of generator bus which is normally monitored by PMU
[3].
The proposed algorithm is divided into two parts, the
first part is calculating the system COI through data
collection from PMUs placed on generators bus. The
second part, a fast learning algorithm is used to predict

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
127

the fore coming values for COI and consequently

determine system stability measure.

of angular position c in x-y coordinates and so as to

the predictor position.

3.1 Center of Inertia COI

The COI is calculated in [5] as follows:

The x-y components of the predictor will be:

COI =

i Hi
i =1
N

Hi

px = cos (p)

py= sin (p)

(1)

and the center of inertia angular position will be

represented by:

i =1

Where i is the internal generator rotor angle, H i is

the respective generator inertia time constant and N is
the total number of generators of the system areas.

cx = cos (c)
cy = sin (c)

Equations 4, and 5 will be substituted into the coarse

tuning objective function:

Min T f ( c p ) + [( cx px ) 2 + ( cy py ) 2 ]

T f ,

(6)
where, T f is a rough estimate of the final prediction

For phase angle measurements in an area i in

interconnected power system, we have the approximate
center of inertia COI angle reference:
N

ij Pji
j =1

j =1

Where

(5)

For each sampling instant of PMU the center of inertia

angular position x and y components will be updated.

As the machine inertia is directly proportional to the

real power output, the weight factor of generator inertia
time constant can be replaced by generator real power
P.

i
COI
=

(4)

time obtained from the equation:

(2)
P ji

p c + T f ( p c ) = 0

i
j

P denotes the real power generation schedule

at generating plant j =1,...N in area i.

c , p

Increasing the number of phase angle measurements, in

each area of interconnected power system, accuracy of
i
computation can be increased.
COI
For the entire interconnected power system, the overall
center of inertia c can be calculated as follows:

velocities respectively, and are represented by:

d
c = c = c ,
dt

d p
p = p =
dt
and and are weighting functions.

c =

i
Pi
COI
i =1

Pi

(7)

are the center of inertia and predictor

(8)

The minimization of the objective function is obtained

by differentiation of equation (6). The predicted center
of inertia angular velocity p in y-direction is driven

(3)

i =1

through the following equation:

A1 3py + A2 2py + A3 py + A4 = 0

3.2 Prediction Algorithm

The proposed prediction algorithm is an
adaptation of a proven robotic ball-catching algorithm
and has been applied to power system instability
prediction as in [4].

(9)

where, A1, A2, A3, and A4 are functions of cx , cy , cx ,

cy , , and which are known.

The prediction algorithm is divided into two parts

described in [6]:

The predicted center of inertia angular velocity

p in

x-direction is driven through manipulating equations (6)

and (7)
( px cx )
px = cx +
( py cy )
(10)
( py cy )

The coarse tuning which is a tracking stage of

interconnection center of inertia, c , and
The fine-tuning which is the extended prediction of
center of inertia, c .

The optimum predicted center of inertia angular

velocity vector can then be obtained, through Solving
equation (9) for py and equation (10) for px .

A. Coarse tuning
The actual calculated center of inertia angular
For each sampling instant the difference between the
position is represented by c and the predictor is
measured center of inertia c and the predicted center
presented [Link]
p. The
algorithm
requiresonthe
projectionand Electrical
International
Conference
Environment
Engineering, Karpacz, Poland, May 10-13, 2009
128

of inertia p is calculated and compared to a

predetermined tolerance TOL. If the difference between
the actual and predicted center of inertia is less or equal
than pre-specified tolerance [TOL (p - c )], the
algorithm is switching into the fine-tuning stage of
prediction.
B. Fine tuning
The purpose of this stage is fine tune the
predicted center of inertia value p after a period of Tf
regarding the p trajectory generated from the coarse
tuning process. Taylor series expansion has been proven
as a good estimator for unknown data. It is used in this
stage to fine tune the values obtained from the previous
coarse tuning stage.
Angular center of inertia velocity varies continuously
and shows a smooth change because of the large inertia
of turbine-generator combination of interconnected area
power system. In order to fine tune the predicted
angular velocity p , and canter of inertia angle p, a
function extrapolating the three measured points c(0),
c(1), and c(2) is defined using Taylor series
expansion given by:

p (T f ) = c ( 2 ) + 1 (T f 2 ) + 2 (T f 1 )(T f 2 )
(11)
where; 0, 1 and 2 are the last three time stamps for
PMU measurements.
Tf = the prediction period in sec.
0 = ( c ( 1 ) c ( 0 )) /( 1 0 )
1 = ( c ( 2 ) c ( 1 )) /( 2 1 )
2 = ( 1 0 ) /( 2 0 )
(12)
Integrating equation (11), the function predicting the
center of inertia p will be given by:

p (T f ) = p (t ) + c ( 0 )

4.1.1 Stable conditions

A. Generators COI angles prediction
The system is simulated using Matlab
software. The test system contains eleven buses, four
generators, four transformers, two shunt capacitors, and
two loads. The technical data of the system are obtained
from [5].
A three phase fault is initiated at 0.5 s at bus nine of the
studied system. The duration of the fault is 0.1 s and is
cleared by opening the line 9-8 at 0.6 s. Simulation
results showed that the system is stable.
The application of the prediction algorithm showed that
the COI generator angles can be predicted for 500 ms
with an acceptable accuracy.
Figure 2 shows the actual and predicted COI angles for
the four system generators. It is obvious that the
predicted values are very close to the actual COI
measured values.
The prediction errors between the actual and predicted
COI of the four generator buses 1, 2, 3, and 4 are also
calculated. Figure 3 shows the prediction error of the
four COI and the error is very small compared to the
actual value of COI angles.

(13)

By using the algorithm above, area said to be losing

synchronism can be predicted without any prior system
configuration and exhausting system calculations. It will
be shown by application that the proposed system can
predict inter foreseen area instability in a time window
permit the application of proper preventive control with
acceptable prediction error.
4. APPLICATION TO INTERCONNECTED
POWER SYSTEM

0.5
Actual and Predicted Relative Generator Angles, rad.

Tf

Fig. 1 Single line diagram of simple test system.

0.4
0.3
0.2
0.1
0
-0.1
-0.2
-0.3
-0.4

4.1 Simple Test System

The applicability of the proposed algorithm to
a wide area power system is first studied with a simple
system of two areas. The basic topology is depicted in
Fig. 1 [5].

The system contains eleven buses and two areas,

connected by a weak tie between buses 7 and 9. The left
half of the system is identified as area 1 while the right
half is identified as area 2.

50

100

150

200
250
300
Time, msec.

350

400

450

500

Fig. 2 Actual and predicted COI angles for a prediction time of 500
ms.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
129

0.03

0.02

4.1.2 Unstable conditions

When applying a three phase fault at bus 6 at a
fault time 0.5 s and clearing the fault 0.11 s later
removing line 6-7 at 0.61 s, the system gone unstable.
The proposed algorithm is applied to the unstable
condition and the prediction results are given in figure
6. Area A1 COI angle decreases in an exponential
manner to reach a value of -22.5 radians.

Error, rad.

0.01

-0.01

50

100

150

200

250

300

350

400

450

500

Time, ms.

Fig. 3 Prediction error for COI bus angles.

B. Two area systems:

The previously 4 generators test system is
divided into two area as given in [5]. Area 1 contains
buses 1, 2, 5, 6, and 7, while area A2 contains buses 3,
4, 9, 10, and 11. The two areas are connected through a
tie line between bus 7 and 9. The same fault and
clearing conditions are applied to the two areas system.
The proposed algorithm is applied to the test system to
predict the two areas COI angles. Figure 4 shows the
simulation result for 500 ms. Results verify that the COI
angles of the two areas can be predicted with an
acceptable accuracy. From figure 5, the prediction error
for the COI angles is around 0.01 radian through most
of the prediction period.
Actual and predicted Area1 and Area 2 COI angles, rad.

0.2
0.15

A1
0.1
0.05

-0.05
-0.1

-10

A2

-0.15

50

100

150

200

250

300

350

400

450

500

Fig. 4 Actual and predicted COI angles for area 1 and area 2
0.1

-15

Actual COI A1

-20

50

100

150

200

250
Time, s

300

350

400

450

0.06
E A1

Fig. 6 Actual and Predicted COI angle for two areas network.
0.1

0.05

-0.05

Area A 2

-0.1
Area A1

-0.15

-0.2

50

100

150

200

250
TIme, s

300

350

400

450

0
-0.02
E A2

-0.06
-0.08

50

500

Figure 7 shows the error between actual and predicted

COI angles and it was found that the error value
increases for unstable area A1 to reach about -0.24
radian.

0.02

-0.04

500

4.2 IEEE 50 Gen-145 bus Test System

The investigation is, further extended to a
larger test system. The IEEE 50 generator test system
given by [7] is considered. The system data are given in
table 1. Full system details can be found in [7].

0.08

0.04

Predicted COI A 1

Fig. 7 Error between actual and predicted center of inertia of two areas
system

Time, ms

Error of COI angle for area 1, and area2, rad.

-5

-0.25

-0.2

Actual and Predicted COI A 2

0

-25

Error between actual and predicted COI, rad.

-0.03

Actual and predicted COI angles of area A1 and A2, rad.

-0.02

100

150

200

250

Time, ms.

300

350

400

450

500

Table 1: IEEE 50 Generator Test System Data

No. of Generators
50
No. of Buses
145
No. of Transmission Lines
401
No. of Transformers
52
No. of PV Buses
49
No. of Loads
64
No. of Shunt Reactance
97
System Frequency
50 Hz
Slack Bus
100

Fig. 5 Error between actual and predicted COI angle for area 1 and
area 2
8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
130

400

Considering the wide area control and the need to divide

the network into interconnected control areas; the
system is divided into two areas as shown in table 2[7].
Table 2: IEEE 50 Generator Areas
Area No.

To investigate the effectiveness of the proposed

algorithm, a three-phase fault was applied at bus 7.
Fault was initiated at 0.1 second and cleared at 0.208
second by tripping line 6-7. This fault created an
unstable condition for the system as generator at bus
104 (Generator #2) accelerating increasingly as shown
in figure 8. In the meanwhile, the other 49 generator
maintain equilibrium.

Generator Bus No.

Actual and predicted COI angles, rad

350
GEN2 Bus104

300

actual

250

115, 116, 130, 131, 132, 134, 135, 136, 137, 139, 140,
141, 142, 143, 144, 145
60, 67, 79, 80, 82, 89, 90, 91, 93, 94, 95, 96, 97, 98, 99,
100, 101, 102, 103, 104, 105, 106, 108, 109, 110, 111,
112, 117, 118, 119, 121, 122, 124, 128

1
2

200

Figure 11 shows the COI angle of the two area IEEE-50

generator system. For the fault conditions described
above area 2 COI accelerate and the area is going to an
unstable zone. Area 1 is stable as the variation of the
inter area COI angle is approximately steady.

predicted

150
100
50
0

Actual and predicted COI angles for 49 GEN

-50

50

100

150

200

250
300
Time, ms

350

400

6
450

500
5

Fig. 8 Actual and predicted generator angle referred to COI at fault

conditions

4
Area COI angle, rad.

The prediction error is calculated and plotted to show

the deviation from the actual COI angle. Figure 9 shows
that the prediction error of generator 2 at bus 104 is
increasing till a maximum value of 3.25 radians.

2.5
COI prediction Error, rad

COI angle of area A 1

3
GEN2 error

-1

2
1.5

50

0.5
0
GEN 1 and 3 through 50 error

-1

150

200

250
300
Time, ms

350

400

450

500

50

100

150

200

250
300
Time, ms

350

400

450

500

Fig. 9 Prediction error in radians

Predicting the two areas COI angles, the results shown

that a precise approach to the values which have been
calculated before.
Figure 12 shows the actual and predicted COI angles of
the IEEE-50 generators two control areas. It is clear that
the predicted values converged too close to the actual
value.

Separating generators (1, and 3 through 50) prediction

error; figure 10 shows that prediction error clustered
around zero with maximum value of 0.08 radians.

4
Areas COI angles, rad.

0.1
Prediction error of GEN 1, 3 through 50

0.08
0.06
0.04
0.02

Actual and predicted COI angles of area A 2

1
Actual and predicted COI angles of area A 1

0
0

-0.02
-0.04

-1

-0.06

50

100

150

200

250
300
Time, ms

350

400

450

500

Fig. 12 Actual and predicted COI angles for 500 ms prediction period.

-0.08
-0.1

100

Fig. 11 The distribution of two areas COI angle along simulation time.

-0.5

COI angle prediction error, rad.

3.5

-1.5

COI angle of area A 2

50

100

150

200

250
300
Time, ms

350

400

450

500

The accuracy of the predicted values is explained

calculating the error between the predicted and the

Fig. 10 Prediction error of stable 49 generators

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
131

actual COI angles. Figure 13 shows that the prediction

error does not exceed 0.3 radians.
0.5
0.4
Prediction error for A1 and A2, rad

COI Prediction error A2

0.3
0.2
0.1
0
-0.1
COI Prediction error A1

-0.2
-0.3
-0.4

50

100

150

200

250
300
Time, ms

350

400

450

500

Fig. 13 Prediction error for 500 ms prediction period.

5. CONCLUSIONS
The proposed algorithm is adopted from a proven robot
control system and the technology of PMU [2, 3, 6].
The algorithm shows a good accuracy regarding the
actual COI angle prediction. No prior knowledge of
system state is required. PMU measurements can predict
the system stability state for 500 ms in advance. Wide
area monitoring and control can be achieved on-line
using the proposed algorithm. Severe contingencies can
be prevented at the right time as foreseen system state
can be predicted in advance. Further investigation on
wide area control, such as generator tripping and load
shedding, using the proposed method is in progress to
proof the applicability of the algorithm.
6. REFERENCES
[1] D. Hu, and V. Venkatasubramanian, "New wide-area algorithms
for detecting angle instability using synchrophasors," in Proc. of the
Western Protective Relay Conf., Spokane, WA, USA, 2006, pp 1-37.
[2] G. Karady, A. Daoud, and M. Mohamed, On-Line transient
stability enhancement using multi-agent technique, in Proc. of IEEE
WM02, 2002, pp. 893-899.
[3] C. Taylor, D. Erickson, K. Martin, R. Wilson, and V.
Venkatasubramanian, "WACS-Wide-area stability and voltage control
system," in Proc. of the IEEE, Vol. 93, No. 5, May 2005, pp. 892-906
[4] G. Karady, A. Daoud, M. Mohamed and I. Amin: Fast learning
algorithm for synchronous generator instability prediction, Intl
Journal of Engineering Intelligent Systems, Vol. 12, No. 2, June 2004,
pp 67-71.
[5] P. Kundur, Power System Stability and Control, EPRI, McGrawHill, Inc., USA, 1994.
[6] B. Hove, and J-J Slotine, "Experiments in robotic catching," in
Proc. of IEEE Int'l Conf. In Robotics and Automation 1995, Vol. 1,
1995, pp. 380-385
[7] Washington State University, IEEE test case archive, Available:
[Link]

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
132

Some aspects of technical, economical and wind

simplified analysis for 18 MW wind farm in
Poland
Waldemar DOLEGA
Institute of Electrical Power Engineering;
The Wroclaw University of Technology
Wybrzeze Wyspianskiego 27
Wroclaw, Poland
Abstract- In this paper, simplified analysis of different conditions
of wind power expansion in Poland is shown. This analysis is
made on basis of simplified computational example for 18 MW
wind farm in Poland from the point of view different groups of
parameters. Four groups of conditions: wind, geographical,
technical and economical are taken into consideration.

I.

energy sources in general amounted to 5,8% [4,9]. Share of

wind generation in total country energy consumption in 2006
amounted to only 0,16 % [4,9].

INTRODUCTION

Renewable energy sources as: wind, hydro, solar, biomass,

biogas, geothermal and sea waves are of more and more
fundamental importance in electricity production in countries
of the European Union. This situation also ivolves Poland.
From renewable sources, the quickest and the greatest
development concerns wind energy. Last years in the world
just belongs to wind power energy. Now, there are wind
turbines set with a total capacity of over 94000 MW in the
world, at the territory of the Europe over 64900 MW [8]. Only
in 2008, at the territory of the European Union countries there
were in total wind turbines with capacity of 8484 MW set,
which means 15% increase of installed capacity in comparison
to the previous year [8].
Germany is still leader of the sector in the extent of market
size, where the total installed capacity exceeds 23903 MW, the
second place in the extent of installed capacity belongs to
Spain - over 16754 MW [8]. Moreover, in six countries of the
European Union the total installed capacity exceeds 2 GW in
Italy (3736 MW), France (3404 MW), Great Britain (3241
MW), Denmark (3180 MW), Portugal (2882 MW) and Holland
(2225 MW) [8].
Wind energy in Poland is still in its early phase in
comparison with other countries of the European Union. At
countrys territory there are 36 wind power plants (only 11
wind farms) with a total capacity of 442 MW [9]. Average
capacity of a set turbine amounts to about 1,52 MW. Capacity
installed in wind energy -per capita is 0,0037 kW, and per km2
of land area 0,45 kW [9]. Wind energy density in Poland is
one of the lowest in the Europe. A number of wind farms is
very small in composition with fact that there are profitable
wind conditions in 30 % area of Poland (Fig.1).
In 2006 total energy production in wind parks in Poland
amounted to 388,4 GWh, which means that share of energy
production from wind in production of energy in renewable

Figure 1. Wind resources in Poland ( /colour localisation/; grey - very

profitable /north ends of Poland/; light grey - profitable; medium grey
enough profitable; dark grey unfavourable /south part of Poland/; black
very unfavourable) [10]

In term 2004-2006, it was a dynamic increase of number

and capacity of wind farms (almost 300%). It was installed
183,9 MW in wind farms installed power in this time [9].
A relatively quick development of renewables in Poland
(especially wind farms) is connected
with special
consideration paid to renewables at the national level. It is
determined in The Energy Law [1] and Decrees (Ordinances)
of Ministers in charge of economic affairs [2,3]. This success
was made mainly thanks to support given to renewables.
Support given to renewables in Poland are represented by:
green certificate payment (market based), subvention to
investment, additional benefits and limited fees for connection
renewables to the grid [1-3].
According to The Energy Law and appropriate decrees (e.g.
Ordinance of Ministry of Economy from 19.12.2005 about
renewable energy purchase obligation [3]) the total balance of
electricity in the gross national energy consumption in 2010
must contain 10.4 % share of renewable energy (and
respectively: in 2009 8.7%). It means that any energy
enterprise whose activity consists in electricity generation or
trade in electricity, which sells the electricity to the final

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
133

customers connected to the grid on the territory of Poland are

obliged to purchase electricity from renewables 8.7% in 2009
and 10.4 % in 2010 [3].
Additionally, polish government plans for 2020 concern
installing 13600 MW in wind energy and 20 % share of wind
generation in the national energy consumption [4]. It means
that in 2009-2020 increase of capacity that is needed amounts
to at least 13200 MW, at the same time connecting about 1200
MW a year.
Presently, this subject matter are specially important because
of fact, wind energy in Poland is planned to develop very
dynamic and in the near period is planed to build many wind
farms [9,10].

12 units about 1,5 MW each (NEG Micon 1500/72,

Nordex S70/1500);
9 units about 2 MW nominal power each (Bonus
2000/76, NEG Micon 2000/72, Vestas V66 2000/66).
Fourth group of consider conditions were economical factors
as: wind turbine price, installation cost of turbine, total
investment cost of turbine, operational & maintenance cost,
price per 1 kWh produced electricity, price of green certificate
per 1 MWh produced electricity, inflation rate and rate of
discount. All incurred costs and economical calculations were
realised in EUR currency. Wind turbine prices was taken from
offer specification of producers [12,13]. Installation cost of
turbine was qualified as 30% wind turbine price [11].
Operational & maintenance cost was considered on level 1.5%
II. DESCRIPTION OF PROJECT
wind turbine price yearly [11]. Period of exploitation was
The main aim of this project was a realisation of qualified as 25 years [5]. Rate of discount was qualified as 8 %
computational example of 18 MW wind farm in Poland from [6,10]. Real rate of return was also qualified on level 8%. For
point of view different groups of parameters. Four groups of simplification, price per kWh produced electricity was taken in
conditions: wind, geographical, technical, and economical were calculations as constant in time on level 0.03, 0.04 and 0.05
taken into consideration.
EUR/kWh. Price of green certificate per 1 MWh produced
First group of conditions were contained wind regime electricity was qualified as 64.21 EUR/MWh [7]. For
(consistent with mesoscale) as: yearly wind speed, specific economical calculations all prizes, costs and counters
yield and Weibull distribution. Yearly wind speed on level: 4, (EUR/PLN, USD/PLN, EUR/USD, EUR/DKK) were taken as
4.5, 5, 5.5 and 6 m/s, profitable for wind power expansion in average level in 2007. Investment was qualified as totally
Poland, was considered in analysis. This range of values of realised from own financial sources. Quantity of subsidies was
wind speeds is fited to conditions which are occured in Poland calculated as the aim to obtain of investment profitability.
(Fig.1.). Wind speed was qualified at a height - 30 m.p.g [10].
Economical calculations were realised with reference to
Simplification as wind regime for temperature of air - 15C, hypothetical wind farm about 18 MW nominal power. For
pressure of air - 101.3 kPa /specific weight 1.225 kg/m3/. simulation, this farm is included from: 30 units about 600 kW
Wind was modelled to use Weibull distribution with shape nominal power each, 24 units about 750 kW each, 18 units
parameter equal 2. This model was a standard for polish wind about 1 MW each, 12 units about 1.5 MW each and 9 units
conditions [10]. Wind speeds were related to geographical about 2 MW nominal power each. All units in the wind farm
regions in Poland.
were produced by the same producer.
Second group of consider conditions were geographical
III. WIND FARM CALCULATION
conditions as: localisation of wind farm and surface class.
Localisation of wind farm was qualified through connection
In frames this project two main groups of calculations were
with yearly wind speed. Surface class (roughness) was realised. These calculations were made with using two
considered according to Danish elaboration on level: 0 computer programs elaborated by author.
(surface of water), 1 (open fields with single buildings), 2
First group were calculations of year energy output for
(open fields with few buildings /distant about 500 m/) [5,11].
specified wind turbines, wind speeds, surface classes and
Third group of
consider conditions were technical Weibull shape parameter. In frames these calculations power
conditions as: unit of wind turbine, producer of wind turbine,
input, power output, energy output and capacity factor were
rotor area, wind turbine power curve, characterised power of
calculated.
unit of wind turbine. In the frames of investigations some
Technical and economical calculations are related to five
units of leader producers from Denmark are considered [11analysed
configurations of wind farm, 15 basic units,
13]. These producers are: BONUS, NEG MICON, NORDEX,
profitable
wind
conditions (yearly wind speed: 4, 4.5, 5, 5.5
VESTAS and WIND WORLD. Offer specification of turbine
and
6
m/s)
and
different
surface classes suitable for localisation
for each producers is taken into considerations [12,13].
of
wind
farm
(surface
class:
0,1 and 2). The main parameters of
For project simulation of wind power plant about 18 MW
technical
and
economical
estimate
were: yearly total energy
nominal power were used:
output
from
analysed
wind
farm,
capacity
factor of wind farm,
30 units about 600 kW nominal power each (Bonus
total
investment
costs,
total
income
per
year
(sale of electricity,
600/44 Mk IV, NEG Micon 600/43, Nordex N43/600,
sale
of
green
certificates,
inclusive),
cost
of
1 kWh produced
Vestas V44 600/44 and Wind World 600/42);
electricity
from
analysed
wind
farm.
Electricity
price per kWh
24 units about 750 kW each (NEG Micon 750/44, Wind
on
assumed
levels
are
determined
as
a
limit
of
profitability,
World 750/48);
according
to
polish
conditions
[10].
18 units about 1 MW each (NEG Micon 1000/60, Bonus
1000/54, Nordex N54/1000);

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
134

Process of simulations and analysis was divided into two

stages:
selection of
the best solution
within given
configuration of wind farm,
selection of the best solution for wind farm.
IV. ANALYSIS OF WIND FARM
Conducted simulations and analysis for individual
configurations of wind farm about 18 MW nominal power let
to select the best variant for each configuration. Selection was
based on three criterions: max. yearly energy output, min. cost
of 1kWh produced electricity and max. total income per year
from sale of electricity and green certificates. The best
solutions for each configurations of wind farm are presented in
table 1. The best composition of wind farm was selected of
these configuration according to listed earlier criterions.
TABLE 1.
The best solutions inside individual configuration
Composition of
wind farm

Producer of
turbine

Type of
turbine

30 x 600 kW
24 x 750 kW
18 x 1MW
12 x 1.5MW
9 x 2 MW

Nordex
NEG Micon
NEG Micon
Nordex
Bonus

N43/600
750/48
1000/60
S70/1500
2000/76

Nominal
power of
turbine
600 kW
750 kW
1 MW
1.5 MW
2 MW

less electricity and total inclusive income per year decreases

also 17 % i.e. 502113 EUR (tables 2, 6). Producing of 1 kWh
electricity costs 0.02 EUR more (table 7).
Significant differences between solutions occur for yearly
wind speed on level 6 m/s. The farm included 30 units Nordex
N43/600 produces 14 % less electricity than farm included 18
units NEG Micon 1000/60 (table 2). Differences are smaller for
other configurations of wind farm.
Similar correlations concern capacity factors for wind farm.
Maximum values of capacity factors obtain for yearly wind
speed on level 6 m/s. Than wind turbine NEG Micon 1000/60
has capacity factor on level 27 %, however turbines Nordex
N43/600 and NEG Micon 750/48 have 23 % (table 3).
TABLE 3.
Comparison of capacity factor (%) for analysed 18 MW wind farm
different yearly wind speeds (surface class = 0)
Composition Producer Type of
Yearly wind speed, m/s
of wind
of turbine turbine
4
4.5
5
5.5
farm
30 x 600
Nordex
N43/600
8
11
15
19
kW
24 x 750
NEG
750/48
7
11
14
19
kW
Micon
18 x 1MW
NEG
1000/60
9
13
18
22
Micon
12 x 1,5MW Nordex S70/1500
8
12
16
21
9 x 2 MW
Bonus
2000/76
8
11
15
20

for

6
23
23
27
26
24

Analysis of solutions for optimal localisation of wind farm

(surface class 0) from point of view total energy output for
different yearly wind speeds leads to conclusion that the best
solution is a wind farm included 18 units NEG Micon 1000/60.
Yearly wind speed has the greatest influence on total energy
output of wind farm. A higher yearly wind speed in place of
wind farm localisation means larger productivity of wind farm.
Even slight deviation from assumed yearly wind speed results
in differences in electricity production on level a few dozen or
so, and even a few dozen percent. Differences between five
considered solutions with reference to total energy output are
presented in table 2. Solution with 1 MW units is the best for
each analysed yearly wind speed, but in principle solutions
with units about largest nominal power are the best.

Simplified economical analysis leads to conclusion that the

wind farm included 18 units NEG Micon 1000/60 has the best
economical parameters.
Total inclusive income for wind farm is obtained from sale
of electricity and green certificates. It depends mainly on
quantity of produced electricity as well as unit price for sale of
electricity (1 kWh) and unit price for sale of green certificate
(1 MWh). For example, wind farm included 1 MW units
brings in total inclusive income per year on level 4416713
EUR for yearly wind speed 6 m/s, however it decreases 65% to
level 1534230 EUR for yearly wind speed 4 m/s (table 6).
Influence of unit price for sale of electricity on total income
of wind farm for example yearly wind speed 5 m/s is
presented in table 4.

TABLE 2.
Comparison of total energy output per year (MWh) for analysed 18 MW
wind farm for different yearly wind speeds (surface class = 0)
Composition
Yearly wind speed, m/s
of wind farm
4
4.5
5
5.5
6
30 x 600 kW 11838.9 16803.6 22914.0 29406.3 36280.5
24 x 750 kW 11421.0 16750.9 22842.1 29694.7 36928.1
18 x 1MW 14722.5 20968.3 27660.4 34798.5 42382.8
12 x 1.5MW 12954.4 19026.8 25908.9 33600.6 41292.3
9 x 2 MW 12168.6 17537.1 23621.4 30779.4 37937.3

TABLE 4.
Comparison of total income per year (EUR) obtained from electricity sale
in function of price for 1 kW produced electricity
(yearly wind speed = 5 m/s, surface class = 0)
Composition Producer of
Type of
Price for 1kWh, EUR
of wind farm
turbine
turbine
0.03
0.04
0.05
30x600 kW
Nordex
N43/600
687420 916560 1145700
NEG
24 x 750 kW
750/48
685264 913685 1142106
Micon
NEG
18x1 MW
1000/60
829811 1106415 1383018
Micon
12x1.5 MW
Nordex
S70/1500 777266 1036355 1295444
9 x 2 MW
Bonus
2000/76
708641 944855 1181068

For example, wind farm included 1 MW units produces

27660 MWh electricity for yearly wind speed on level 5 m/s
and it brings in total inclusive income per year on level
2882489 EUR for price for sale of 1 kW produced electricity
on level 0.04 EUR (tables 2, 6). The cost of 1 kWh produced
electricity from analysed wind farm is equal 0.12 EUR (table7).
This solution is on a limit of profitability, thanks to significant
income from sale of green certificates. Solution with 750 kW
units is the most disadvantageous. Such farm produces 17 %

Influence of unit price for sale of green certificate on total

income of wind farm for different yearly wind speeds is
presented in table 5. For example, wind farm included 1 MW
units brings in total income from sale of green certificates per
year on level 2721400 EUR for yearly wind speed 6 m/s,
however it decreases 65% to level 945332 EUR for yearly
wind speed 4 m/s (table 5).

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
135

TABLE 5.
Comparison of toal income per year (EUR) obtained from sale of green
certificates for produced electricity for different yearly wind speeds
(surface class = 0)
Composition
Yearly wind speed, m/s
of wind
4
4.5
5
5.5
6
farm
30 x 600kW 760176 1078959 1471308 1888179 2329571
24 x 750kW 733342 1075575 1466691 1906697 2371153
18 x 1MW 945332 1346375 1776074 2234412 2721400
12 x 1.5MW 831802 1221711 1663610 2157495 2651379
9 x 2 MW 781346 1126057 1516730 1976345 2435954

Profitability of wind farm depends strictly on cost of 1 kWh

produced electricity. It is connected with yearly wind speed,
higher speed means smaller cost (table 6). Cost of 1 kWh
produced electricity should be smaller than price per kWh
produced electricity. Analysed solutions arent effective
economicly even for the highest assumed level of unit price
for sale of electricity (0.05 EUR) and yearly wind speed 6 m/s
(table7). Effectiveness is obtained thanks to significant income
from sale of green certificates.
Wind farm included 1 MW units has the smallest cost of
1 kWh produced electricity. However the largest cost is for
wind farm included 750 kW units (table 7).
TABLE 6.
Comparison of total income per year (EUR) obtained from electricity sale
and sale of green certificates for different yearly wind speeds
(price for sale of 1 kW produced electricity - 0,04 EUR surface class = 0)
Composition
Yearly wind speed, m/s
of wind farm
4
4.5
5
5.5
6
30 x 600 kW 1233732 1751103 2387868 3064431 3780791
24 x 750 kW 1190184 1745611 2380376 3094487 3848276
18 x 1MW 1534230 2185109 2882489 3626353 4416713
12 x 1.5MW 1349980 1982785 2699965 3501518 4303070
9 x 2 MW 1268089 1827540 2461585 3207519 3953448
TABLE 7.
Comparison of electricity cost per kWh for analysed wind farm for different
yearly wind speeds
(price for sale of 1 kW produced electricity - 0,04 EUR, surface class = 0)
Composition Producer Type of
Yearly wind speed, m/s
of wind
of turbine turbine
4
4.5
5
5.5
6
farm
30 x 600
Nordex
N43/600 0.27 0.19 0.14 0.11 0.09
kW
24 x 750
NEG
750/48
0.28 0.19 0.14 0.11 0.09
kW
Micon
18 x 1MW
NEG
1000/60 0.22 0.15 0.12 0.09 0.075
Micon
12 x 1,5MW Nordex S70/1500 0.25 0.17 0.13 0.09 0.08
9 x 2 MW
Bonus
2000/76 0.26 0.18 0.135 0.10 0.08

V. CONCLUSIONS
A quick development of renewables (especially wind farms)
in Poland in last years is closely related to good law and
economic solutions.
The Energy Law [1] with Ordinances of Ministry of
Economy [2,3] are key law acts in Poland for development of
renewables.
Support given to renewables in Poland constitutes market
based green certificate payment, subvention to investment,

additional benefits and

limited fees
for
connection
renewables to the grid.
Green certificate payment (market based) and subvention to
investment are key support means given to renewables in
Poland.
Green certificate payment is the best solution for
development of renewables applied in Poland.
Energy enterprises which generates electricity from
renewables in Poland have two sources of income - from sale
of physical electricity and - from sale of property rigths
connecting with certificates of origin (green certificates).
Realised investigations enable to qualify a localisation of
wind farm about 18 MW nominal power for different profitable
yearly wind speeds, different surface classes and optimum
selection of unit for wind farm.
Yearly wind speed has the greatest influence on total energy
output of wind farm and electricity cost per kWh produced by
it. Growth of yearly wind speed causes considerable growth of
total energy output and considerable decrease electricity cost
per kWh.
Surface class has considerable influence on total energy
output and electricity cost per kWh.
Type of units applied in wind farm has a great influence on
total energy output and electricity cost per kWh.
The best solution for analysed 18 MW wind power plant is
farm consisted of 18 units NEG Micon 1000/60.
At present market mechanisms wind power farm
investments are profitable for good localisations. The main
causes of wind farm profitability are market based mechanism
of green certificates and subvention to investment from
different national and union funds.
REFERENCES
[1]

Act of 10 April.1997 The Energy Law (Journal of Laws of 2003

No.153, Item 1504, No.203, Item 1966, of 2004 No.29, Item 257, No 34,
Item 293, No.91, Item 875, No 96, Item 959, of 2005 No.62, Item 552).
/in polish/
[2] Ordinance of Ministry of Economy from 4.05.2007 about detailed
conditions of power system functions (Journal of Laws of 2007 No.93,
Item 623). /in polish/
[3] Ordinance of Ministry of Economy from 19.12.2005 about renewable
energy purchase obligation (Journal of Laws of 2005 No.261, Item 2187).
/in polish/
[4] Energy policy of Poland up to 2030 Project formulated by Ministry of
Economy. Warsaw, september 2007. /in polish/
[5] T. Burton, D. Sharpe, N. Jenkins, E. Bossanyi: Wind Energy Handbook,
John Wiley and Sons Ltd. Chichester, England, 2001.
[6] J.F. Manwell, [Link], A.L. Rogers: Wind Energy Explained:
Theory, Design and Application, John Wiley and Sons Ltd. Chichester,
England, 2002.
www pages
[7] URE (Energy Regulatory Office) [Link]
[8] EWEA (European Wind Energy Association) [Link]
[9] PWES (Polish Wind Energy Society) [Link]
[10] Wind Plants News Bulletin [Link]
/in polish/
[11] Know
how:
Danish
Wind
Industry
Association,
[Link]/en/tour.
[12] Wind energy: The World of Renewable Energy Industry, [Link].
[13] Wind turbine index: The International Portal for Wind Energy - Branch,
[Link].

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
136

The aspects of combustion and co-combustion

biomass
Grzegorz Maliga a, Amar Patil b
a

Phds student of Departament of Mechanical and Power Engineering, Wrocaw University of Technology
[Link]@[Link]
b
MScs student of Departament of Computer Science, Wrocaw University of Technology
amar_ec006@[Link]
Abstract

Growing world population, increasing energy use,

emission carbon dioxide (CO2) and climate changing are
the reasons for searching alternative sources of power.
Recently we are vested with numerously alternative
sources of energy but biomass is considered as a most
powerfull and easiest to access.
The bigest problems with burning raw biomass are: first
the calorificvalue of biomass is average half compare to
coal, second problem is a big contents of moisture for
example in coal 10% of weight is water but biomass
include even 35% of water.
For these reasons in huge power plants complete
displacing stone fossil and lignite by biomass is not
workable. Coal-fired furnaces which are used in power
plants are not adapted to burn raw biomass, hence in
powerplants co-combustion biomass and coal for power
production is most efficiency because this process does
not require additional energy. Precious advantage of cocombutstion is that technical range of change the fuel
feed system is less. Preparing the fuel mixture based on
blending biomass with coal on storehaus or in coal mills.
The most efficient dose biomass to coal is about 10%
weight.
Otherwise in small domestic heat furnace burning a raw
biomass with high efficience is possible for example
wood pellets furnaces are popularly used in onefamily
houses. These kind of combustion occure, because the
furnaces are specific adapted. The calorific value of
wood pellets is near 19 MJ/kg which is compare to coal.
Wood pellets are assembled for example numerous
from energy willow or other kinds of waste wood and
preparing the pellets form reqires energy addition.
In this paper I demonstrate two ways of burning
biomass: combustion and co-combustion. I try to show
that selecting the way of burning is depended by size of
furnace.

I. INTRODUCTION
Today human life is becoming very mobility and
more comfortable, because it is representative aspect
of modern civilisation. In different point of view,
more comfortable life influences on level of energy
consumption. We are aware that increase energy

using is the main reason of growing the greenhouse

effect and warming the atmosphere which symptom
as the unpredictable changes in climate. In this issue
only one effect emission of carbon dioxide and other
greenhouse gases by the combustion of fossil is
predicted. Currently 90% energy which population
use becomse from fossil supply. It dimishes coal and
oil amount. From these reasons we require
alternatives to fossil sources of energy, which can be
renewable.
In table I are presented the sources of renewable,
which were used in Germany in 2007. It shows that
biomass is the bigest share of renewable energy and it
amounts 48,6%, where 37.9% was used as a source
for heating and 10.7% for producing the electricity.

TABLE I
SHARE OF RENEWABLE ENERGY CONSUMPTION IN
GERMANY 2007
Nr.
1.
2.
3.
4.
5.
6.
7.
8.

Groups of renewable energy

sources
Biomass (heating)
Biomass (electricity)
Biofuels
Hydropower
Wind energy
Solarthermal energy
Geothermal energy
Photovoltaic energy

Share
%
37.9
10.7
20.0
9.3
17.8
1.7
1.0
1.7

Potencial of using biomass as a source of energy is

huge compare to the other kind of renewable fuels.
The second improtant subject of debate is race of
energy, which can be produced from biomass. The
most share is heating and electricity and these are
most usefull in todays life. In oder to produce
heating or electricity from biomass it is necaessary to
use the burning process, which converts chemical
capacity of biomass to thermally kind of energy.
Biomass could be burn as a main fuel which is
considered as a combustion or the alternative is cocombustion where biomass is only added as a share to
fossil fuel for example in 10% amount.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
137

II. BIOMASS
Biomass is regarded as material of organic origin. In
this group count not only plants but also animals
excrements, plant components, paper, cleulose,
organic waste and vegetable oil.
There is a lot of method to use these various organic
materials as a source of energy. In some technical
solutions raw biomass can be turn into liquid or
gaseous form before converting to other kind of
energy. Raw organic materials can be burn in furnace
to prepare heat. The second way is use the
fermantation process by anaerobic digester to yield
biogas, which can be latter burn. Third variant is
using thermochemical gasification to convert biomass
into synthetic gas. But this paper focused only on
aspects of combustion and co-combustion.
A. Sources of biomass
Important source of biomass is the forestry. For
example in Germany 80% of timber, which is annual
growth become felled and used. Felling and preparing
timber consequence in a big amount of waste timber,
which left in the forest. Timber waste is no useful as a
construction timber or for celulose production, so it is
adequante oppotunity to use it as a biomass for
energy production.
Second source of biomass is agriculture.
Currently for production the same amount of food
farmers need less and less land. From this reason
cultivation of energy crops can increase and in the
future it could be great energy potential. Today the
residues materials like a straw are used prevalent and
are converted for heat in specific furnance. Also the
fast growing groups of trees and grass are tested to
increase these efficiency, because in the future these
could be the main sources of biomas. Production
energy crops and plants by farmers becoming greater
because using new method of cutivation being
economically.
Third source of biomas is food industry. For
example in process of preparing fruits or vegetables
produce a lot of green waste, wchich can be used as a
biomass. But the bigest disadvantage of using this
kind of biomass is large amount of moisture. Before
using green waste as a fuel for furnace, it should be
dry or mixing with other king of organic materials,
which contain less amount of water.

B. Porcess of growing biomass and carbon dioxide

neutral.
The basic of plants growing reason is the
photosynthesis occurence. In this reaction carbon
dioxide with water are transfrometed into oxygen and
carbohydrates. The oxygen is released into
environment. The main force of photosynthesis is
solar energy. The main conclusion could be that for
the fact of being the sun is possible to use biomass.
The efficiency of producing biomad is depended on
naturaly differences as the soil and climate
opportunieties. It is calculated [1] that every year in
the World is growing 400 milion tones of biomass.
This amount of material contains 30001018 J of
energy. For example human use every year about
4001018J, where only 451018 J is producing from
biomass. These numbers shows that in the future
biomass could be use more intensively as a source of
energy.
The carbon dioxide aspect in growing plants is
important role, because it is fixed by plants, which are
called as a carbon dioxide sink. This phenomenon
influences on earths climate balance. Also raw fossil
materials are carbon dioxide sink, but fixing existed
milions years before. Now the raw fossil contain large
amount of carbon and these are easy used as a fuels in
combustion process. Comnining carbon with oxygen
produces carbon dioxide which is main reason of
greenhouse effect.
Nature biomass growing and using as a source of
energy make that the concentration of carbon dioxide
is constant and the cycle is closed, but when the fossil
are used carbon dioxide concentration rise.
TABLE II
AMOUNTS OF CO2 EMISSIONS DEPENDING ON THE
ENERRY SOURCE

Energy source

Fuel oil
Natural gas
Wood pellets
Firewood

CO2
Annual CO2
emission
emission
(kg/MWh)
(kg/a)
342
228
68
8.8

5472
3648
1088
141

CO2 savings
compared
with fuel oil
(kg/a)
0
1824=33%
4384=80%
5331=97%

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
138

III CO-COMBUSTION BIOMASS

In huge powerplants, where using power coal as

a basic fuel is common, because from the energy
poiny of view, coal characterises two features:
effortlessness to purchase and the calorific value is
very high level such as 22 MJkg-1. Compare biomass
to power coal the calorific value is half lower. Second
important issue is fuel feed system, which was
designed and adapted only for fossil coal. Replacing
whole of fuel feed system requires a huge amount of
capital expenditure and it is unfounded from
economical point of view. The third issue is problem
with adapting the steam boilers to new fuel such as
biomass. Majority of steam boilers, which are using
in huge power plants are adjusted to coal dust. Coal
dust is fed to steam boiler as a powder after milling
process. Granularity of the fossil dust is depended on
standards of quality and for example in polish power
plant industry are using coal dust where R90 value is
30% and R200 is about 3%. It means that for biomass
obtaining similar value of granularity is not possible.
Another problem with biomass is level of moisture,
which is double compare to coal. Bigger biomass
particles and high value of moisture need longer time
of burning in boiler combustion chamber compare to
coal dust. Different is also the bulk density of
biomass because the value is only 0.25-0.30 kgm3,
wchich is very low compare to coal powder. Another
physical properties, temperatue of burning and
amount of ash reason that the furnaces should be
fixed apart to different kind of biomass. During
designing process of co-combustion biomass with
coal is important to consider the share of added
biomass, because it influences directly on parameters
of
working the steam boiler such as: steam
temerature, exhaust temperature, amount and
chemical composition of ash or burning waste and
slaging the heating surface, which operating
conditions is important from heating change point of
view. Basing on the results of research [1] cocombustion the biomass, showed that the influence
share of biomass does not increase emission of
nitrogen oxides (NOx) , sulphur oxides (SOx) and
carbon oxides. For a variety of biomass only adding
a rape straw increases contents of sulphur oxides (
SOx) in exhaust gases. Besides adding the biomass to
major fuels does not influence on slag and ash
properties. Experiments [1] allow for proof that only
addind straw changes rheology properties of slag,
because especially the corn straw contains high
quantity of chlorine. Content chlorine in fuels increas
the flow ability of slag which is negativ phenomenon,
bacause when the opportunity of slag to flow is
higher, it consecuences of becoming overgrown the
heater surface by slag. When the layer of slag
increase on the boilers surface, it will decrease the
heat exchange between exthaus gas and working
medium, such as water or steam, because slag

cahracterises high heat resistance. For example when

the heat is exchanging in the way of conduction the
efficiency is defined by Furriers law, which is
describes by equation (1):

Q=

A T

[1]

where Q is the heat flux, W; the thermal

conductivity, W(m2K)-1; l the thickness of layer heat
exchange, m; A the heat exchange surphase, m2; T
the temperature difference, K.
Basic on the equation (1) is able to conclude that
for the material such as slag, the thermal conductivity
is less, so when layer of slag become higher, the heat
flux decrease.
In huge power plants when the mainly fuel is
coal, the overall efficency of thermal power plant
cycle is about 27-36% and the energy consumption is
about 9-13.7 MJ/kWh. When the clear biomass is
burning in the same steam bioler the efficiency of
thermal cycle becomes lower than before and is
about 18-27% and the energy consumption is higher
compare to clear fossil and is about 13.7- 21.1
The differtent between coal and
MJ/kWh [2].
biomass efficiency of thermal cycle cause of moisture
value of biomass [1].Tests in power plants proofed
that co-combustion becomes most efiiciency way to
use biomass by steam boiler. When the share of
added biomass is about 7-10% the boiler efficiency is
lower only 1% compare to clear coal dust.
In thermal-electric power station of Gorzw the
influence of added biomass to steam boiler type OP140 was researched by [1]. The purposes of
investigation was to find the most efficience share
value for energetic willow, which is added as a
biomass to coal dust. In thermal-electric power plant
30% of energy production is converting into electra
energy form and about 42% converting into thermal
form ( as hot water and technology steam).
As a alternative fuel added to coal is energetic
willow, because the opportunity of access huge
amount of this kind of biomass is highest compare to
other, because the consumption biomass in power
plant achives high value. The local market is possible
to deliver enough amount of willows timber, this fact
becomes willow timber the most attractive as a
biomass. Calorific value of energetic willow is
comparable to other kind of timber and it is about
10 MJ/kg. Moisture contents in raw willows timber
determine about 45%. Based on assumption, the
maximum of moisture contecnts should be such as
30%, from this reason before adding, the willow has
to by dryed. The chemical compopsition of willow is
similar to other timber is presented in table III [1]
compare to coal fuel.
Costs of production the raw willows timber is about
10 $/t, after milling and drying the total costs
bacomes about 20 $/t, which is still profitable.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
139

TABLE III
CHEMICAL AND THERMAL PROPERTIES OF COAL AND
BIOMASS

C
H
O
N
S
Volatile matter
Calofific value

Coal dust

Biomass

57.39 [%]
3.726 [%]
6.012 [%]
0.893 [%]
0.63 [%]
26 [%]
22,4 [MJ/kg]

35 [%]
7 [%]
23.03 [%]
1.4 [%]
0.07 [%]
80 [%]
13.5 [MJ/kg]

Analysis co-combustion willows timber in steam

boiler OP-140 showed that process of burning two
types of fuel together is more beneficial than burning
coal and timber apart, because coal stabilizes burning
process, which allows to use biomass with variable
contents of moisture. Whereas added biomass
dimishes emission powders, ashes and gases such as
NOx and SOx. Timber contains less incombustible
parts than coal. The value is about 0.5-3% for timber
and almost 12% for coal. When fuel characterises
high value of incombustible parts, it causes that
exhaust gas contests fixed higher value of powder and
ashes. Timber includes only 0.3% of nitrogen, which
is two times less compare to coal. Contents of sulphur
is near 0.05% where for coal it becomes 0.8%.
In thermal electric power station biomas is added to
coal before milling process. The tests [1], which was
realized by coal pulverizer type fan
MWk-16,
showed that the degree of share added biomas till
30% does not influence in negative way on working
of coal pulverizer. Blending biomass with coal on
coal store before milling decrease the operating costs,
because this solution does not need install special
kind of feed system for adding the biomass. During
milling coals ad biomass together taked place not
only milling and mixing but also drying the biomass,
which is profitable.
During burning process mixed coal with biomass in
steam boiler type OP-140 the temperature of exhaust
charactrizes almost constant value. The analysis of
working the combustion chamber showed that adding
biomass to coal in value about 10 % does not
influence in negative way on work parameters.
Process of co-combustion biomass with coal share in
range about 7-12% does not change the parameters of
steam boiler and the work efficiency is comaparable
when the same boiler works based only on coal fuel.

IV CONBUSTION BIOMASS IN SMAL-SCALE

BOILERS
Biomass contains less energy in proportion to
their weight compare to raw fossil materials, from
this reason it is better to use them where they occure.
Heat and electric energy can be produced in small or
medium plants which are decentralized. These kind of
boilers are used in blocks, flats, schools, swimming
pools or lockal heating networks. The small heat
plants which are supplied by biomass can attain
power of beteen 500 kW to 30 MW. They mostly use
woodchips from thinningor wood pellets from
[Link] kind of materials are cheep and ensure
as a alternative fuel, because many municipalities use
woodchips from their own woodland.
For example in rural areas there are more and more
used smal-scale household biomass boilers supplied
with biomass. These kind of burners are generaly
instaled for log and pellets. They get more benefits,
because for example pellet boilers run fully
automaticly.
Log boilers can attain heat capacity from five to
several hundred kilowatts. The disadvantage of log
burners is that the wood has to be put into furnace
still by hand. The combustion is very efficient and
cleaner, because a lot of them fixed lambda probe,
which controll there is enough oxygen for completly
burning process and less air polution such as carbon
monoxide.
Pellets are in pressed form object 6 or 8 mm
thick and about 10 to 30 mm long. Pellets calorific
value is about 18 MJ/kg ( 2 kg of pellets are eqivalent
of 1 litre of oil fuel). They are also very easy to
transport and take up less space to storage. Pellet
boiler are equiped with screw conveyor anf fan,
which can adjust the lewel of required heat.

REFERENCES
[1]

[2]

[3]
[4]
[5]
[5]
[6]

D. Futrak, M. Ostap, Wspspalanie biomasy w

niskoemisyjnym palenisku wirowym na przykadzie prb w
EC Gorzw, Confernce of Co-Combustion Biomass and
Renewable Fuels in Steam Boilers, 12-14 may, 2004.
Z. Bis, W. Nowak, Authothermal Upgrading of Biomass and
Wastes for Co-combustion in Power Systems, Confernce of
Co-Combustion Biomass and Renewable Fuels in Steam
Boilers, 12-14 may, 2004.
[Link]
[Link]
[Link]
A. Iwanick, Biomass for Steam and Power, Harris Group Inc.
C. Owczarek, H. J. Krautz, S. Griebe, Cycloid Combustion- a
New Concept for Industrial Combustion with Low Calorific
Fuels, Chair of power plant technology, Brandenburg
Technical University of Cottbus, Germany

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
140

Control and Stability Analysis of Doubly Fed

Induction during Voltage Sags
*D. Kairous, *B. Belmadani, **M. Benghanem
*Faculty of Science and Sciences Engineering, UHBC, Po Box 151, Chlef 02000, Algeria,
** Electrical engineering Depart. USTO. , Po Box 1505, El Mnaouar, Oran, Algeria
Mail: kdjilali@[Link]
Abstract- Large wind turbines are often equipped with
Doubly Fed Induction Generator -DFIG-.
Today the penetration level of wind power on the
network is high. The power system stability defined as the
ability of the system, for given initial operation condition,
to regain a normal state of equilibrium after being
subjected to a disturbance is the major problem.
In numerous papers the stability problems are pointed
out and there are also many papers describing control
strategies for wind turbine equipped with DFIG.
In this paper we study dynamic behavior of the DFIG
during voltage sag. Also we study the ability of the Control
to remind the stability of the system.
I.

INTRODUCTION

Wind energy is one of the most important and promising

sources of renewable energy due to its clean character and free
availability. Many variable speed winds turbines have been
used with induction generator, permanent magnet and
eventually switched reluctance generators. Nowadays the
market is oriented on the design of high power wind generation
systems based on multi-pole synchronous machine or doubly
fed induction generators (DFIG).
Power system stability has been recognized as an important
problem for secure system operation [1]. The pining of the
market of electricity obligates us to make production of the
generators near their physical limits, for economic reason. It is
thus necessary to evaluate these limits, in particular the risk of
instability of network voltage.
Different control strategies for such kind of generator have
been proposed for active and reactive power control or voltage
and frequency control. Direct field oriented control for active
and reactive powers is presented in [2], [3].
In the proposed paper we present the control of the Active
and reactive Powers by sliding mode. The system has the
ability for simultaneous maximum wind power generation for
large speed range of operation and achieves soft and fast
synchronization to the grid. Power converters and associated
control strategies are simulated using Simulink.
Some simulation results are presented to validate the
theoretical analysis and to show the behaviour and
performances of the proposed structure for power control.

Figure.1. Representation of DFIG-based wind turbine

II. DESCRIPTION OF THE STUDIED SYSTEM

The basic configuration of the whole system is presented in
Fig. 1. The rotor of DFIG is connected to the grid through two
back to back bridge converters. The grid side converter (GSC)
is used to control the DC-link voltage and to keep it constant
regardless to the magnitude and direction of the rotor power.
The DFIG is controlled by the rotor side in order to generate
the optimal active power depending on the wind speed and
turbine characteristics.
III. GRID SIDE CONVERTER CONTROL
The role of the grid side converter is to keep the DC-link
voltage constant regardless to the magnitude and direction of
the rotor power. A vector control approach is used, allowing
independent control of the active and reactive powers flowing
between the supply and the GSC.
Fig.2 shows a simplified representation of the grid connected
converter. AC-side series inductances accounts for transformer
leakage reactance, and series resistances represents inverter and
transformer conduction losses
The model of the system is given by the following equation:

ia
1 0 0 ia
Va Van
d
R
1

ib = 0 1 0 ib + Vb Vbn

dt
L
L
ic
0 0 1 ic
Vc Vcn

(1)

With the Park transformation, the equation system of the

GSC is given by:

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
141

d id

v dr = R i d + L d t s L i q + Vdn

v = R i + L d i q L i + V
q
s
d
qn
qr
dt

The direct axis current is used to regulate the reactive power

and the quadratic axis current is used to regulate the DC-link
voltage. The reference frame is considered oriented along the
stator voltage vector
This method gives possibility to realise independent control
of the active and reactive power between the GSC and the
supply side.
The voltage components in park frame are given at the
output of the PI regulator. The voltage source components and
the powers can be written as:
vd = 0 , vq = V

P = vq iq and Q = v q i d

III. SLIDING MODE CONTROL

(2)

(3)

We will use the sliding Mode Control to control the rotor

currents of the DFIG. The goal is to have a more stability in a
disturbance cases then the PI [Link] write the Park
model of the DFIG [4] :

d dqs

m s . qds
Vdqs =R s Idqs +
dt

V =R I + d dqr m .
r
qdr
dqr s dqr
dt

dqs =Ls Idqs +M.I dqr

dqr =Ls Idqr +[Link]

(6)

(7)

By neglecting the stator resistance of the stator we can write:

By neglecting converter losses, we have:

v dc i dc = v q i q
vdcC

dv dc
, C
= idc i m
dt
dvdc
= P Pm
dt

Vds = 0

Vqs s ds

(4)

(8)

dr =L r I dr + L . Vqs

s
s

=L I + M V
qr
ds
qr r
L s .s

(5)

Based on the above relations, the GSC control diagram can be

easily deduced as presented in Fig.2

(9)

The general model of the machine is given by:

d
M
[Link] [Link] + Vqs sqs + ds

Lss
dt

Vds
dqs

R
.
a

R
.
c
L
I
+
V
+

V
s
qs
s r qr
ds s ds

L
dt

s s

qs =

Vdr

dIdr
M
M
[Link] + Vqs [Link] r LrIqr + Vds +Lr
Lss
Lss
dt
Vqr

dIqr

M
M
Rb
Vds [Link] +r LrIdr + Vqs +Lr
r. LrIqr +
Lss
Lss
dt

(10)

1
1
With: a=
, b=
,
Ls
Lr

M
c=
LsLr

The state model is put in the following form:

.

X =f(x,t)+g(x,t)U dq
With:
Figure.2. Blocks diagram of the Grid side converter control.

U dq = Vds Vqs Vdr Vqr

(11)

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
142

1
0

g(x,t)= 0

0
1
0
Lr
0

0
0

1
Lr

M
[Link] +[Link] + Vqs +sqs

Lss

[Link] +[Link] + Vds sds

Lss

f(x,t)=

1
M
M
[Link] + Vqs +[Link] +r LrIqr + Vds
Lss
Lss

Lr

1 R .bLI + M V +R.c LI + M V
Lr r r qr Lss ds r ds r r dr Lss qs

Figure.3. Sliding mode control of the DFIG

powers are keeping at their nominal values.

The DFIG used has a nominal power of 20kw.
We make a voltage sag of 50% during 200 ms (Between
t=0,45s and t=0,65s).
The results in the Fig.4 (figures of the currents, powers and
the torque) prove the validity of proposed model, and the good
stability of the system.

The slip surfaces in the Park reference are defined to control

the rotor currents.
They are given by the following equations [5]:

d = ( I drref I dr )

q = ( I qrref I qr )
Where

(12)

V dr and V qr are the two control vectors of, they force

the trajectory of
surfaces dq = 0 .

the

system

to

converge

-a-

towards

The vector U eqdq is obtained by imposing dq = 0

f ( x, t ) + g ( x, t ) Vdq = 0

(13)

M
M
[Link] + Vqs +[Link] +r LrIqr + Vds
Lss
Lss

Ueqdq =

[Link] + M V +R.c LI + M V
r r qr L ds r qs r r dr L qs
s s
s s

-b-

(14)
The Fig. 3 presents the global diagram for the DFIG with
sliding mode control.
V. SIMULATION RESULTS
In this section we present the simulation results of the system
(Fig.1). The rotor voltage the stator voltage, the speed and the
powers are keeping at their nominal values.

-c-

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
143

NOMENCLATURE

J , C fr

: Inertia and viscous friction

: Mechanical speed
turbine : the turbine angular frequency
Vd , Vq , I d , I q

stator and rotor Voltages, currents

ds , qs , dr , qr :

Two-phase stator and rotor flux

ids , iqs , idr , iqr Two-phase stator and rotor currents

-d-

I dms , I qms Two-phase measured rotor currents

Cr , C e Prime mover and electromagnetic torque

Rs , Rr : Per phase stator and rotor resistance
M : Magnetising inductance
Ls , Lr : Total cyclic stator and rotor inductances
s , r : Pulsation
g : Generator slip
i d , iq Two-phase grid side converter currents
v d , vq
-eFigure.4. Dynamic behaviour of the DFIG during Voltage sag of 50 %
a- Stator Current, b- Rotor current
c- Active Power. d- Reactive Power. e- Torque

VI. CONCLUSION
In this paper a doubly fed induction generator connected to
the grid was presented.
The aim of the paper was to develop the decoupled d-q
vector control technique of DFIG supply by a back-to-back
PWM converter in the rotor side. The control strategy contains
two control levels: DFIG control level (control of active and
reactive power using sliding mode approach) and the control of
the grid side converter.
The mathematical model of the system (DFIG-ConvertersGrid) has been implemented in MATLAB & Simulink.
The main goal of the grid side converter control is to keep the
dc-link voltage constant by balancing the real power on the
machine side and on the grid side converter, and to compensate
reactive power of the DFIG to get the unity power factor.
The control system is applied to the rotating reference frame
fixed on the gap flux of the generator. It has been proved that
this control system can control the active and reactive power
independently and stably.
The simulation results show that we can make decoupling
between active and reactive power and in the same time have a
good stability in a case of voltage sag.
Also, the DC-link voltage was kept constant to the reference
value. The torque has the same dynamic like the active power.

Two-phase network voltages

vdc , i dc : Dc-bus voltage and current delivered by the grid side converter
P , Q : Active power delivered by the grid side converter and
im , Pm : Current and Active power delivered to the rotor side converter
C : Capacitance of the DCink
R , L Resistance and Inductance of the grid
Ps , Pr : Stator active power and rotor active power
Q s Qr : Stator reactive power and rotor reactive power
v ds , v qs , v dr , vqr :Two-phase stator and rot
REFERENCES
[1] P. Kundur, J. Paserba, Definition and Classification of
Power System Stability, IEEE Transaction on Power
Systems. Vol. 19, N2, May 2004.
[2] J R. Pena, R.C. Clark, G.M. Asher, Doubly Fed
Induction Generator using Back-to-Back PWM Converters
and its Application to Variable-Speed Wind-Energy
generation, IEE Proc-Elect. Power Appl, Vol. 143, N5.
380-387. 1996
[3] L. Mihet-popa, I. BOLDEA, Control Strategies for Large
Wind Turbine Application, Journal of Electrical
Engineering, Vol. 7, Edition 3rd, ISSN 1582-4594. 2006.
[4] P.C. Krause, Analysis of Electrical Machinery, McGrawHill, New york, 1994.
[5] B. Toufik, Commande Structure Variable et Etude de
lIntegration dEolienne base de Machine asynchrone
Double Alimentation sur le Rseau Electrique, PhD thesis
(in
French),
University
of
NANTES,
2007

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
144

Application and Comparison of LQR and Robust Sliding Mode

Controllers to Improve Power System Stability
S. Saffar Shamshirgar1
Sahar_Saffar@[Link]

M. Golkhah2

H. Rahmati3

[Link]@[Link]

M. A Nekoui4

h.rahmati1@[Link]

Manekoui@[Link]

[Link]. student, Islamic Azad University of Sciences and Researches, Tehran, Iran.
2
PhD student, Ecole Polytechnique de Montreal, Montreal, Canada.
3
[Link]. student, K. N. Toosi University of Technology, Tehran, Iran
4
Assistant Professor, Engineering Faculty, Islamic Azad University of Sciences and Researches, Tehran, Iran.

Abstract Heffron-Philips model has had great popularity to

simulate mono or multi machine power systems. Since the stability of
this model is related to operating point of synchronous generator(s),
many efforts have been made in research papers to design robust and
reliable controllers to ensure the stability of the system. A typical
mono-machine power system is presented in this paper to make a
comparison between behaviors of traditional power system stabilizer
and LQR based pss. Characteristics and fundamental concepts of
each controller are stated. At the termination of the paper, Sliding
Mode method has been engaged so that the power system can be
stable in uncertain condition. The results of LQR base and sliding
mode based PSSs have been compared under uncertainty.
Keywords Heffron-Philips model, pss, stability, LQR.

I.

INTRODUCTION

Connecting small generation and distribution systems

together makes better stability against small disturbances but
when a serious problem occurred for one of the connected
systems, it also affects the others. These connections have
been done between even different countries for economic
reasons. Modern power systems have many characteristics
such as far distances of consumption from generation
declining the power system stability. In order to provide
stability for such modern systems, different control systems
have been designed and applied in power plants and power
transmission lines. Drum level control, Governor, and
Automatic Voltage Regulator are such control systems which
are applied in power plants. Governor and AVR have bound
capability to damp transient state oscillations. Therefore
another control system called power system stabilizer, pss, is
applied to improve the stability. This additional loop can be
applied for both governor and AVR control loops. Applying
pss to governor makes many quick variations in the position of
mechanical fuel gate making it injured. Hence it is
recommended to apply pss to AVR controller [1].
Static Var Compensator, SVC, synchronous condenser, tap
changer, and FACTS devices are control systems in power
transmission lines.
To simulate and control a generator, it is important to have
a model. State space model is a well-known for this aim. The
complete model for a synchronous machine has 7 orders. It is
efficient to reduce the degree of the system as well as
linearization in order to simplify the model. Heffron-Philips

model is such a linear model with 3 orders. This model is

considered to apply controllers in this paper.
II. CASE STUDY

Figure 1. The power system anticipated as case study

Figure (1) indicates a power system as our case study. It

includes a power plant comprising four same units. The power
plant is connected to a transmission line by a transformer. The
line conveys the power to an infinitive bus.
Electrical parameters presented on table (1) are pertaining
to normal regime operation of the system [1]. The values are
in per unit system.
Table 1. Initial values for the power system

Generator constants

H=3.5, D=0.5, T'do=8.0

Xd=1.81, Xq=1.76, X'd=0.3

Exciter constants

KA=10

Line and transformer

constants

XL1=0.5, XL2=0.93,
XT=0.15

Electrical parameters

P=0.9, Q=0.3, Vt=1.0<36o

VB=0.995<0o

III. HEFFRON-PHILIPS MODEL

Heffron-Philips model for generator is linear model with 3
degrees in state space. Figure (2) shows this model with AVR
loop.
The state space matrices corresponding to the presented
model in Figure (2) can be formulated as equation (1).
Where [x1 x2 x3] = [J K e'q] and [u1 u2] =[TM uE].
It is easy to calculate all the parameters of equation (1) by
using table (1).

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
145

The response of speed deviations of the rotor after

removing the second line with and without pss loop is shown
in figure (4).

Figure 2. Heffron-Philips model with the AVR controller

x1

k1
J

x2 =

k4
k 3T ' do

x3

1
J

kA
T ' ldo

D
J
k5k A
T ' do

1
k 3T ' do

u1
u2

k2
J
k6k A
T ' do

x1
x2
x3

Figure 4. Speed deviations of rotor with and without presence of

pss after the fault

(1)

Moreover, responses of power angle and deviations of

terminal voltage of generator with and without pss can be seen
in figure (5) and figure (6) respectively.

IV. PSS CONTROL LOOP

Suppose a condition in which the second transmission line
is removed under a fault [1]. The aim is to study rotor's speed
oscillations and apply a pss to improve the stability resulting
in decrement of speed deviations around the steady state
value. Steady state values in such a condition can be
calculated as equation (2).

Vdo = 0.6836,

Vqo = 0.7298

I do = 0.8432,

I qo = 0.4518

= 79.13,

E fdo = 2.395

(2)

Figure 5. Power angle without and with presence of pss after the
fault

Using these values, coefficients k1 through k6 have below

values.

k1 = 0.7643,

k 2 = 0.8649

k 3 = 0.323,

k 4 = 1.4187

k 5 = 0.12

k 6 = 0.3

(3)

Input signal of pss is speed of rotor and the output is a

bounded voltage adding to the generator excitation. A block
diagram for a regular pss is indicated in figure (3).

Figure 3. Block diagram of a regular pss

Figure 6. Terminal voltage of the generator with and without
presence of pss after the fault

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
146

It is proved from figures (4), (5), and (6) that pss has been
able to improve stability of the system actually by shifting the
eigen values of the matrix A of state space equations to the left
areas of s-plane. Since stability concept for a generator and
related equipments is angular stability of a power system, you
can say that pss extends the angular stability limits of a power
system. However, pss can exacerbate instabilities of a power
system. It is reasonable because pss is designed to operate
around an operating point so it can improve the stability under
small disturbances. However, when a large fault occurred, pss
can lose synchronism of the generator by providing surplus
excitation field. In addition, you may think of tuning a pss to
extend more the angular stability but there is indeed a
limitation to tune the parameters of pss since its output voltage
reaches to saturation levels determined by the actual
limitations. If the parameters of pss were selected in a way to
have small amplitudes in state variables, the output voltage of
pss becomes larger. On the other hand if they were designed in
order to have small output voltage of pss, the amplitudes of
state variables become larger. Therefore a compromise should
be employed.
V. LQR BASED STATE FEEDBACK
Closed loop poles of a linear control system can be
emplaced in desired places of s-plane using state feedback and
observer control system design. Also its poles can be chosen
by choosing appropriate observer gain. Response speed and
estimation error dynamics can be defined by choosing closed
loop poles. However, optimal selection of closed loop poles is
really hard for industrial systems and real processes. Albeit an
unstable system can be stabilized by applying states
feedbacks, but linear optimal control systems should be
engaged for below reasons.
The first reason is that it is hard to find appropriate closed
loop poles in which the desired behavior of the system is
satisfied. Selecting closed loop pole with great negative real
parts makes the dynamic response of the system to be quick
while the control effort to be greater than permissible levels. If
selection of the closed loop poles makes saturation of control
signals, dynamic behavior of the system will not be as same as
the desired behavior even it may become unstable.
The second reason is noise which specially occurs in the
systems with high gains.
Therefore optimal selection of closed loop poles will lead
to a trade-off between speed of dynamic response and control
effort.
Suppose the linear system in state space given by equation
(4).

Where Q is a non-negative definite matrix which is called

weighting matrix. On the other hand, R is a positive definite
weighting matrix. Selection of values of these matrixes
determines the dynamic speed of the controller as well as
amplitudes of state variables and control signals. For example
if R is selected small while Q is selected large, more stability
will be attained with large control efforts.
Solving an optimal control problem means finding a u(t)
by which equation (5) is minimized. Linear Quadratic
Regulator, LQR, can solve this problem [2] (see equation (6)).

AT X + XA XBR 1 B T X + Q = 0

(6)

After calculating X from algebraic equation (6), k and u can

be calculated from equations (7).

k = R 1 BT X
u = kx

(7 )

Indeed, LQR is an optimal pole assignment and integral J

defines circumstance of assigning closed loop poles as an
optimizing criterion.
For our case study R and Q matrixes have been selected as
equations (8).
It has been tried to have small control effort and equal
importance to the state variables. It is important to note that
there will be guarantied gain and phase margins by LQR. In
fact it is not required to verify stability of the system after
designing state feedback by LQR because it must be stable.

R=

0.5 0
0 0.5

1 0 0
, Q= 0 1 0
0 0 1

(8)

VI. OBSERVER DESIGN

If A and C matrixes were observable in equation (4), then
the observer can be designed as equation (9) [3].

x = Ax + Bu + L(Cx

y)

(9)

Where x denotes estimations of states x. x = x x is

estimation error and it should reach to zero. x can satisfy
equation (10).

x = ( A + LC ) x

(10)

Behavior of estimation error is related on the eigen values of

matrix A+LC. For the mentioned case study L has been
selected in a way to have -2, -5, and -11 as eigen values of
A+LC (equation (11)).
L= [1696 -17.2 -2039.9]
(11)

x(t ) = Ax(t ) + Bu (t )
y = Cx

VII. LQR BASED POWER SYSTEM STABILIZER

(4)

The purpose is to stabilize the system so that all the state

variables become zero by any initial value in maximum speed.
There are many criteria to do that. Quadratic integral equation
is a well-known such a criterion [2].

J=

1
T
T
( x Q x +u Ru ) d t
20

(5)

The designed observer estimates all the three state

variables. Then the estimated states are fed to the inputs by k
which is designed using LQR.
Simulation results prove that state variables have
significantly smaller amplitudes than those of pss while
having small control signals too. Figure (7) indicates the speed
deviations while figure (8) and figure (9) indicate the control
signals. It can be found from figure (7) that LQR based pss has
significant stabilizing characteristics in comparison with the

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
147

traditional pss. Also figure (8) and figure (9) prove that control
signals are much less in LQR based pss than the traditional
one.

Figure 10. speed deviations of both controller after the heavy

fault
Figure 7. Comparison of speed deviation in LQR based pss and
regular pss

Figure 11. First control signal of the LQR base controller after
the heavy fault
Figure 8. First control signal comparison between the two
controllers

Figure 12. Second control signal of the LQR base controller after
the heavy fault
Figure 9. Second control signal comparison between the two
controllers

Another simulation has been made to ensure that the LQR

based pss has a good reliability. It has been supposed that a
three phase short circuit at the received end of the second line
has been occurred without removing the line. Figure (10)
indicates speed variations of both controllers at the same
condition. LQR based controller is still stable while traditional

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
148

pss is unable to stabilize the power system. This result proves

the extent gain and phase margins of the LQR solution. Figure
(11) and figure (12) indicate that the LQR based controller is
subjected to have greater control effort to remain stable.
VIII. SLIDING MODE BASED POWER SYSTEM STABILIZER
Reference [4] has proposed sliding mode to design a
controller in order to have robust state covariance assignment.
This method is briefly illustrated and is employed to have a
robust controller for the stability problem of this paper.
The main aim of this controller is to reach the answer
paths, trajectories, to a surface. Then the state variables should
be tended to asymptotic stability through this surface. To
achieve this aim the surface should be made attractive so that
it can attract the paths. Hence the controller should be
designed to tend the paths to the surface. The controller should
have switching ability to keep the path by which the state
variables have been reached to the desired surface. Different
surfaces and control rules can be determined regarding the
system model and control desires. This paper has engaged the
proposed method in [4] to determine the plane.
Let plane S to be as (11).
t

S (t ) = Cx(t )

(CA + CBG)x( )d

(11)

Figures (15), and (16) present the speed deviations in LQR

and sliding mode methods respectively under uncertain model.
These results prove that the sliding mode controller is robust
since even under uncertain model it has been successful to
tend the state variables to the surface S. Moreover, Figures
(17), and (18) denote the first and second control efforts of
both controller.

Figure 13. Speed deviations response of LGR based pss in certain

model

m1

Where S (t ) = [ S1 (t ) ... S i (t ) ... S m (t )]

, C and G

are constant matrices to be design, C is chosen such that CB is

nonsingular, and G is the control feedback gain matrix which
should be determined so that the state variables can fit the
requirement in the sliding mode [4]. Equation (12) is attained
by derivative of (11).

S (t ) = CBu CBGx

(12)

Note that CD=0. In the sliding mode the states

satisfy S = 0 , then we get the equivalent control as follows

ueq (t) =Gx(t). Equation (13) can be attained for dynamic
sliding mode by replacing this input into system equation.

x = ( A + BG ) x

(13)

k>

If u(t) is anticipated as (14) in which

positive

A , P is a

Figure 14. Speed deviations response of sliding mode based pss

in certain model

constant,

sgn( S (t )) = [sgn( S1 (t )) ...sgn( S m (t ))] , and x denotes

T

second norm of x, then the state of the system will converge

to the sliding mode surface s(t)=0 with probability. We have
chosen

C=

0 01
,
010

= 1 and k=1 so that CD=0 and CB is

nonsingular. Eventually G is selected so that the eigen values

of A+BG are {-3, -6 5i}.
Let compare responses of certain models of LQR and
sliding mode. Figure (13) represents the response of LQR
based pss and Figure (14) is the response of sliding mode
controller. It is proved from these figures that both controllers
provide good stability condition for the power system however
the amplitude of speed deviations in sliding mode is less than
that of LQR controller.

Figure 15. Speed deviations response of LQR based pss in

uncertain model

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
149

Figure 16. Speed deviations response of sliding mode based pss

in uncertain model

unstable while the LQR based pss was able to stabilize the
power system although the control efforts were increased
significantly. This phenomenon can prove the extent gain and
phase margins of LQR based state feedback. Therefore a
designer can apply LQR to design a power system stabilizer
with good gain and phase margins without any worry about
control signals since LQR can compromise between values of
state variables and control efforts. At the last section, slide
mode controller, which supposes to be robust, applied to the
power system to pole placement. In certain model the results
of both LQR and sliding mode were stable and acceptable
taking it into consideration that the state variables have
smaller magnitudes in the case of sliding mode controller.
Under uncertainty, it was found that LQR controller which is
not a robust controller became unstable while sliding mode
controller was able to remain stable proving its robustness.
X. REFERENCES
I.

P. Kundur, "Power System Stability and Control", McGraw-Hill, New

York, 1994.
II. William L. Brogan, "Modern control theory", 3rd ed., Englewood Cliffs,
N.J., Prentice Hall, 1991.
III. K. Zhou, J. C. Doyle, and K. Glover, "Robust and Optimal Control.
Prentice Hall", New Jersey, 1996.
IV. Koan-Yuh Chang, Wen-June Wang, "Robust covariance control for
perturbed stochastic multivariable system via variable structure control",
Elsevier, Systems & Control Letters 37 (1999) 323-328.

Figure 17. First and second control signals of LQR based pss in
uncertain model

Figure 18. First and second control signals of sliding mode based
pss in uncertain model

IX. CONCLUSION
Fundamental theory of regular pss, LQR based state
feedback, observer, LQR based pss, and eventually sliding
mode based pss were presented in the paper. Simulation
results for a case study including regular pss and LQR based
one were presented separately after removing a transmission
line from service. LQR based pss yielded better stabilizing
parameters while smaller control efforts than the regular pss.
Moreover, after applying a heavy fault the regular pss got
8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
150

A SURVEY OF SOLAR ENERGY POWER SYSTEMS

Adekunle Babatope AJAYI
Donetsk National Technical University
Bogdana Kmelnitskogo 100
Donetsk 83015, Ukraine

ABSTRACT
Using the direct space SUN RAY energy, the proposed solar
energy power system is to links up Industrial to Domestic
usage (I2D) such that each nations can identify an effective
power generation, which brings development to a developing
nation and also stability to the develop nation in every facet of
her economy. This paper highlights the opportunities for
sustaining the developed nations and the developing nations in
the implementation of Renewable Energy Sources by taking
the Free Gift of Nature (FGN) sun, details some possible
architectural configuration and infrastructural requirements.
Concerns over global warming are causing shifts in energy
policy and energy sourcing around the world. Citizens and
governments are carefully weighing their impacts on the
environment; many begin to convert to clean renewable
energy sources. The European countries are leaders amongst
the continents of the world, as well as a nest for the minds of
young independent thinkers and countries. Exploration of the
renewable energy options will allow any country to
understand its potential for reducing its impact on global
warming, such that it can weigh the costs of Climate Change
with those of upfront economic investment in renewable
energy.

Categories and Subject Descriptors

[Energy Applications]: Resource Potential Investors,
Industrial to Domestic (I2D).

General Terms
Management, Documentation, Generation, Transmission

Keywords
Solar Energy, Free Gift of Nature (sun), Power Generation,
Transmission, Environmental and Economic Importance of
Solar Energy to Industrial and Domestic use.

1. INTRODUCTION
The study of renewable energy sources has been of global
concern to the world, and has led many institutions like the
European Commission and others to undertake research on
sustainable approach to meet the challenges of sustainable
energy generation. Renewable energy is a clean energy system

that has no effect during or after generation on the

environment and which this has led to continuous
improvement on solar energy for a better way of reducing
green house effect in the future. This has also helped
developed and developing countries to take full advantage of
this free gift of nature to promote ecological and social
innovation which will ensure more sustainable economy
growth, environmental conservation and social stability.
A critical challenge that continues to constrain the
advancement of many developing countries is the prevalence
of poverty. In spite of an abundance of solar energy most
developing nations, lack stable power source, due to low
technology advancement, poverty, and poor management of
existence facilities.
It is widely accepted that access to electricity is essential to
any nations economy, and is a requirement for modern
economic and social development. Electricity opens the door
to a host of technologies that promote education, public health,
and economic development, such as emissions-free light,
refrigeration, and communication devices. Without electricity,
communities are unable to participate in the benefits of
modern advances and are left isolated and literally in the dark.
The overall objective of the solar energy power system is
towards a stable power supply for developing and developed
nations of the world. Hence, the sunlight that falls on the
surface of a tropical country in one day contains more than
twice of the energy the country can consumes in a year.

2. SOLAR RESOURCE
Solar resource is a process through which the sun emits energy
at an extremely large and relatively constant rate, 24 hours per
day, 365 days per year. Through this process, energy could be
converted into usable forms on earth; it would be more than
enough to supply the worlds total energy demand. However,
there are many effects of ionospheric scintillation on the
environment that will make the days of the year for solar
generation not to be effective: firstly, the earth intercepts only
a small fraction of the energy that leaves the sun; secondly, the
earth rotates such that a collection device on the earths
surface is exposed to solar energy for only about half of each
24-hour period; and finally, the conditions in the atmosphere,
such as clouds and dust, sometimes significantly reduce the
amount of solar energy reaching the earths surface.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
151

2.2 Climate of Nigeria

In this survey Nigeria will be taken as a case study for
reference.
Located on the western part of Africa, Nigeria weather
primarily features a tropical climate where the season is very
humid and damp. There are two climatic seasons that prevail
within Nigeria, namely the wet and the dry seasons. The
weather of Nigeria is generally quite hot throughout the year,
although there are variations in the climate in certain regions
within the country. The southern part of Nigeria is relatively
more humid and damp than the northern part of the country. In
the southern areas of Nigeria the dry seasons begin from the
month of November and lasts till the month of March. The
northern regions are much drier in nature in comparison to the
southern parts. The dry seasons begin from the month of
October and last till the month of April. There are extreme
weather conditions in the deserts of Sahara. It is scorching hot
in the afternoons and freezing cold during the nighttime.

Climate is the average weather condition of a place over a

given period of years. It is determined primarily by distance
from the coast and secondly by evaluation (Bradley 1995).
The climate of Nigeria is tropical in nature, which is
occasionally subjected to variations, depending on the rainfall.
During summers, major portion of the country comes under
the influence of moisture-laden tropical maritime air.
Temperatures are high throughout the year, averaging from
25 to 28C. In the higher elevations of the Jos Plateau north
central area of Nigeria, temperature is at an average of 22C.
Northern Nigeria experiences greater temperature extremes
than the south. Rainfall varies widely over short distances
from year to year. Nigerias electrical energy consumption in
the year 2001 is 15 106 KWh and the tropic of cancer. Its
climate varies from tropical to subtropical. There are two
major seasons; the dry season lasting from October to March
and rainy season lasting from April to October. In the north, it
is hot and dry, rainy season extends between April and
September. In the south, it is hot and wet season extends
between March and December. From December to March,
there is a long dry season (Ojo 2000). Therefore, the
Temperature range between 320 and 420 humidity is about
95% (Falade 1995).

3. SOLAR POWER
3.1 What is solar power?
Solar power is produced by photovoltaic, or "PV", solar panels
and other devices that capture the energy in sunlight and
convert it to electricity. This electricity can then be fed
directly to a consumer, an electric power grid, or a storage
device. Typically, solar panels are installed on the roof of
residential or domestic buildings, and use the power generated
to meet the owner's energy needs and provide surplus
electricity to the grid. Other applications include heating water
and providing power in areas where electricity connections are
not available, such as on road signs, cellular phone towers, and
satellites.
3.2 Environmental and Economical Importance

Fig. 1: Map of Nigeria

2.1 Yearly Irradiation in Nigeria
The following map displays the quantity of energy that
reached the ground in Nigeria for the whole year. This yearly
irradiation is expressed in kWh/m. This map is computed
from observations made by meteorological satellites from
1985 to 2004. European Commission 2002-2006 and
HelioClim-1 copyright Mines ParisTech / Armines 2001-2006.

Solar energy has both the environmental and the economical

importance to every nation. Hence, the solar power play a key
role in cost effectiveness of any nations economy, create a
direct employment of labor force and to foster the
development of micro industries.
The most important factor driving solar energy system
generation process is whether the energy it produces is
economical. Although there are factors other than economics
that enter into a decision of when to use solar energy; i.e. no
pollution, no greenhouse gas generation, security of the energy
resource etc., design decisions are almost exclusively
dominated by the level of energy cost. This similar economic

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
152

parameter, gives the expected cost of the energy produced by

the solar energy system, averaged over the lifetime of the
system. Hence, solar energy power system is a very clean
energy that if given financial support by the government and
industrialist to reduce the cost of implementing the solar
panels for industrial, commercial and residential consumers to
afford.
Abundant Supply: Solar power could meet today's total
electricity demand by PV systems covering only 0.4% of the
nation in a high-sunlight area such as the Southwest an area
about 100 square miles. These panels, in reality, will be
installed across the country on roofs and other structures close
where it is consumed. Technologies such as PV roof shingles,
windows, and flexible fabrics that are easily and cheaply
integrated into new and existing buildings are emerging.
Secure and Stable Supply: Because solar power is generated
domestically, often at the site where it will be consumed,
prices and supplies are immune to blackouts, international
uncertainty and do not rely on long-distance supply networks.
Cleaner Air: Solar power does not pollute air or water. It

replaces electricity generated from facilities powered by coal,

natural gas and other non-renewable fuels, eliminating threats
to public health such as carbon monoxide, particulate, and
toxic chemical emissions from those facilities. Additionally,
when a solar power replaces electricity from a coal-fired
power plant it also eliminates a potential source of sulfur
emissions - a major component of acid rain.

5. GENERATION OF SOLAR POWER

Solar power generation directly converts solar energy into
electrical energy through a chemical action taking place in
solar cells. These operate based on the photo-voltaic effect,
which develops an electromagnetic force (emf) on absorption
of ionizing radiation from sun.
Solar energy (Photovoltaics), the direct conversion of sunlight
to electricity, is now the fastest growing technology for
electricity generation. Present first generation products use
the same silicon wafers as in microelectronics. Second
generation thin films, now entering the market, have the
potential to greatly improve the economics by eliminating
material costs. Martin Green, one of the worlds foremost
photovoltaic researchers, argues in his book that second
generation photovoltaics will eventually reach its own
material cost constraints, engendering a third generation of
high performance thin films. The book explores, self
consistently, the energy conversion potential of advanced
approaches for improving photovoltaic performance and
outlines possible implementation paths.

Reducing Global Warming: Solar power does not produce

CO2 or any other greenhouse gases, thus helping to reduce the
risk of climate change.
4. HOW SOLAR ENERGY WORKS
The solar energy technology is the absorption of suns energy
directly converted into electrical. Hence, the solar
energy works through the aid of combining solar panel
which is also known as photovoltaic cells, solar cells
and solar panels. The solar panel is basically used in
generating 12Volts DC (Direct Current) and the
number of photovoltaic cells used determines the
amount of electricity to be generated.
The solar panels absorb sun ray and then convert it to provide
DC power to the power center, where a centralized
battery bank, or individual home batteries are charged.
The energy from the battery bank can then be used
directly in DC appliances or can be converted to AC
power with an inverter for use in standard AC
appliances. The solar panels can only work during
daylight hours. The solar production system
compliments each other well because it tends to be
windy on cloudy, rain days and sunny on calm clear
days.

Fig. 2: Solar Power Generation Diagram

6. TRANSMISSION OF SOLAR RAY
When solar ray strikes an object, there are two possible
occurrences: the object may reflect the solar energy, or the
object may absorb it. Most objects are applicable for this
process, to a greater or lesser extent. It is useful knowledge to
understand how different materials transmit, reflect, and
absorb solar radiation. For instance, in the case of a solar cell,
it is important to coat the surface with a material that is a poor
reflectorwe want as much light as possible to enter the cell.
Accordingly, creating comfortable, well-lit homes, schools,
and offices requires an understanding of which building
materials transmit, reflect, and absorb solar radiation.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
153

REFERENCE
[1] Baranoski, G., Ronke, J., Shirley, P., Trondsen, T., and
Bastos, R.
[2] Bradley, D. 1995. Plants point the way to renewable
energy. New Sci. 121 : 8
[3] Distribution of Solar Energy System :
[Link]
%20files/[Link]

Fig. 3: Solar Panel Diagram

7. CONCLUSION
The survey has gathered resource potential of solar energy and
how the solar system development in developed countries and
some developing countries (China, South Africa, Gambia,
Kenya, Tanzania, Uganda and many more) and its
implementation in Nigeria.
The most important factor driving the solar energy system
design process is whether the energy it produces is
economical. Although there are factors other than economics
that enter into a decision of when to use solar energy; i.e. no
pollution, no greenhouse gas generation, security of the energy
resource etc., design decisions are almost exclusively
dominated by the level of energy cost. This similar economic
parameter, gives the expected cost of the energy produced by
the solar energy system, averaged over the lifetime of the
system. Hence, solar energy power system is a very clean
energy that should be given support by the government and
industrialists to reduce the cost of implementing the solar
panels for industrial, commercial and residential consumers to
afford.
ACKNOWLEDGMENT
I am grateful to the Electrical Engineering Program Chair of
Donetsk National Technical University, Ukraine Prof. Georgy
G. ROGOZIN, Adedayo Ademola YUSUFF of Electrical
Department, Tshwane University of Technology, Pretoria,
South Africa for their encouragement and guidance. I also
acknowledge the leadership, support and patience of Michael
Adebayo OMIDIORA of Electrical Department, Helsinki
University
of
Technology,
Espoo,
Finland.
[Link] and Olaniyan Kayode
for his advice and support for making this paper a reality.
(ekaynomics@[Link], [Link]@[Link])
National Planning Commission, Abuja, NIGERIA

[4] Eskridge, R. E., O. A. Alduchov, I. V. Chernykh, Z.

Panmao, A. C. Polansky, and S. R. Doty, A
Comprehensive Aerological Research Data Set (CARDS) :
Rough and Systematic errors, Bull. Amer. Meteor. Soc., 76,
1759 1775, 1995
[5] Eurostat News Releases on the internet
[Link]

[6] Falade, T. S. 1995. Solving Housing Problems

in Lokoja. B. Sc. Research Report. Department of
Architecture, A. B. U. Zaria, Kaduna State, Nigeria.
[7] Florida Solar Energy Center (FSEC) 2006,
Photovoltaics & Distributed Generation,
Photovoltaic Fundamentals. Available at
[Link]
[8] National Renewable Energy Laboratory (NREL), 2006a,
Dynamic Maps, GIS Data, & Analysis Tools, Solar Maps.
Available at [Link]
[9] National Renewable Energy Laboratory (NREL), 2006b,
Photographic Information Exchange Available at
[Link]
[10] Ojo, O. 2000. Fundamentals of physical and Dynamic
Climatology, 1st ed., SEDEC Publ., Lagos, Nigeria
[11] Second EU Sustainable Energy week puts the focus on
Energy Supply. January 28 to February 1, 2008
[Link]
[12] SoDa :: Services for professionals in Solar Energy and
Radiation Maps of Radiation, Irradiance, Irradiation and UV
[Link]
[13] Williams, R. H; and Carl, J. W. 1990. Energy from the
sun. Amer. Sci. J. 43 : 41

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
154

Reducing the Short Circuit Levels in Kuwait

Transmission Network. (A Case Study)
Mahmoud Gilany

Wael Al-Hasawi

College of Technological Studies,

Kuwait

Abstract: Preliminary studies on Kuwait high voltage

transmission system show significant increase in the short circuit
level at some of the grid substations and some generating stations.
This increase results from the growth in the power transmission
systems in size and complexity. ew generating stations are
expected to be added to the system within the next few years. This
paper describes the study analysis performed to evaluate the
available and potential solutions to control SC levels in Kuwait
power system. It also presents a modified planning of the
transmission network in order to fulfill this task.

Keywords: Short circuit current, network splitting, fault

current limiter, power transmission planning.
I. Introduction
The topography of the power system network keeps
changing as the economy of a nation grows and its power
consumption increases. One of the effects due to the growth of
energy generation and the expansion of intermeshed
transmission networks is the increase in short circuit currents
in the power systems. These currents can cause mechanical
and thermal stresses and they may damage the equipment.
Consequently, the reduction and control of short circuit
currents is a must.
The problem in Kuwait started at 2005 when a new
emergency generation units were added to the network as a
quick measure to avoid load shedding during heavy-load
summer season. With the long term plan, new large generating
stations are also expected to be installed. As a result, the short
circuit level has increased and it is expected to highly increase
in the near future.
One solution to reduce the risk of increasing short circuit
levels in the power system is to upgrade the protection
equipment so that the estimated fault currents are kept within
the equipment capacity. This usually involves complete
rebuilding of the substation. Nevertheless, this solution does
not represent a high cost ratio benefit. It is an expensive
solution, especially if primary and secondary breakers with
their associated equipment are numerous. In addition to the
cost, upgrading of the existing substations can be a

complicated process because of the number of equipment to be

substituted, redesigned or tested [1].
Instead of upgrading the network, different techniques
aiming at reducing the SC levels were examined, namely:
1. Change the neutral earth policy
2. Un-grounding cable sheath.
3. Changing some lines from AC to DC.
4. Apply current limiters.
5. Dividing the bus bars.
6. Splitting the network
This paper consolidates the experience of the Kuwait
power utility with these measures. The study conclusions are
presented in the form of power system operation strategy in
order to maintain the fault currents within the permissible
limits.
II. The Case Study
Kuwait transmission system consists of a 275 kV Grid
supplied with three major generating stations with a capacity
of approximately 6 GW as shown in Fig. 1. This network is
connected to a 132 kV system which includes another 4 GW
generating stations. Recently, the power system was connected
via 400 kV tie-lines with Saudi Arabia as part of the Gulf
network. Figure 1 is a simplified diagram of the 275 Kuwait
power transmission systems.
Each bus in the Kuwait network is connected to 3 to 4
275/132 kV transformers. The network connected to the
SSUR W 275/132 kV station shown in Fig. 2 is a typical
example. Normally, the loops connected to any bus in the 132
kV networks are isolated from the loops connected to the other
132 kV busses. However, there are normally-open connections
between these buses.
III. Analysis Study
For the purpose of examining the techniques to control the
short circuit, the full Kuwait power network (275 and 132 kV)
was implemented using a simulation program ASPENOneliner V9.7 [2], which is a PC-based short circuit and relay
coordination program widely used by protection engineers.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
155

GS

GBLT W

DWPS

ARDY W

OMAR W

SHKH W

TOWN W

GS

32 km

FRRD W

SSUR W

85 km
SALM X

1
AHMD W

ZSPS

SHUB W

SHUB X

FINT W

SSSM W

OL2K

GS

Fig. 1: The under study 275 kV transmission network

The short circuit level is examined at all the 275 and 132
kV stations. The simulation results show that the short circuit
currents exceed the permissible levels in some specific points.
It reaches 43 kA at some locations in the 132 kV networks
where the nominal rupture capacity is 40 kA. The SC level
also reaches 67 kA at some generating stations in the 275 kV
networks where the rupture capacity is limited to 60 kA. A
typical shape of the output results is shown in Fig. 3.

Fig. 2 Typical part of 132 kV network

Fig. 3 A SLG fault at SURH A station

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
156

This figure shows only a part of the whole network. It

represent a SLG fault applied at station SURH A (132 kV).
The fault current exceeds 37 kA as shown in the figure.
These results did not take into consideration the expected
new generation stations to be installed within the next few
years. If the new generation is considered, the situation will be
much more critical. Therefore, effective activities have to be
planned with aim to reduce the high fault current levels and to
avoid critical situations in the future.
In the remaining of this paper, the measures studied to
limit fault currents are presented.
A. Changing the eutral Earth Policy
Within Kuwait power stations, all the transformers (275
kV and 132kV) are solidly earthed. This type of earthing
system was chosen to avoid any increase in the neutral voltage
during earthing faults. However, the main disadvantage of
such a system is the high magnitude of the short circuit
currents. It was the right time to test the possibility of
changing the earthing policy for some transformers from
solidly earthing to earthing though small resistances.
The simulation results show that with 5 resistance
inserted in the neutral of the local transformers at any
substation, the short circuit level dropped by about 10%. It
dropped to about 15% with 8 resistance. The drawbacks of
this technique include the reduction in the sensitivity of the
overcurrent relays. Practical considerations prevent the use of
higher resistances. The measure has shown to have a limited
effect on the SC reduction target.
B. Changing Some Lines from AC to DC
The idea of changing AC lines to DC lines starts as a result
of the need to increase the power transfer capability of the
existing transmission system without any major new
construction. Conversion of an overhead AC lines to DC lines
means changes in tower head, insulator assemblies and
conductor configuration. There is no need to change the
conductors, tower bodies, their foundations and there is no
intermediate towers added [3, 4].
One of the main features of DC transmission is the fact that
a HVDC connection contributes very little to the short circuit
current. In fact, a back-to-back (BTB) dc connection
functionally appears to be an open circuit from ac impedance
point of view.
It can be seen from the structure of the 132 kV network
under study shown in Fig. 2 that it consists of multiple loops
connected to every 132kV bus-bar. Most of these feeders are
short transmission lines. Consequently, it is neither practical
nor economical to convert all the feeders of each loop to DC in
order to reduce the SC level at this loop. Instead, the
simulations studied the effect of changing some of the "key-

lines" especially those connecting major generation stations in

the 275 kV network. For example, changing the two lines
between DWPS and JABR W (see Fig. 1) from AC to DC will
result in a reduction in the short circuit level at the stations
supplied from JABR W power station by about 30%. For
example, the SC level at station SURH A in Fig. 3 dropped
from 37kA to about 22kA. Similarly, changing the two lines
between ZSPS and JABR W station results in a reduction in
the SC currents at station SURH A from 37 kA to about 23
kA. However, the high cost of the new equipment required
(inverters and converters) and the short distances between
most of the stations in Kuwait make this alternative
inapplicable.
C. Applying Current Limiters
There is a high interest in fault current limiter as it reduces
the expected SC current without affecting the steady state
power flow. There are different techniques to reducing the
short circuit levels using current limiters. The differences
between these techniques include the type and location of
different elements (reactor, thyristor controlled series inductor,
superconductors, etc.) to limit the current. It also varies
according to the network voltage level. The majority of the
current limiters are installed in the tie circuits such as the bus
tie and utility tie circuits in which the current limiters could
perform effectively.
Figure 4 shows a short circuit problem found in the ZSPS
station. This generating station consists of 8x300 steam
generating units plus 4x80 MW gas turbine units. It has a 275
kV bus-bar with four normally-closed sections. The simulation
results show that the SC current reached to a 67.3 kA at the
station bus-bar. The rupture capacity is limited to only 63kA.
One of the most common solutions to the high level SC
currents is the installation of air-core current limiting inductors
(CLI). These CLI are usually installed, although not limited to,
in tie buses and in series with the transmission line [5,6]. With
a 5 reactor added in only one of the three tie-sections of the
station, the short circuit level drops to about 52 kA.
The applicability of adding a reactor for such high voltage
system is facing many problems [5]. Although these series
reactors are the simplest and most economical means of
achieving the reduction in the short circuit levels, stresses due
to the high rate of rise of transient recovery voltages (TRV)
and higher DC offset (X/R ratio) of the short-circuit current
will be experienced by the associated circuit breakers. In
addition, a high DC offset in the fault current may result in
saturation of C.T. cores affecting the performance of the relay
protection schemes [7].
The use of superconducting materials in the development
of FCL looks promising, although at present costs are still too
high. Superconducting materials are used for the reduction of
fault currents, due to their unique characteristic of a fast
transition from zero resistance at rated current to a finite value
with greater current densities. This fast increase in the
resistance results in a fast fault current limitation [8].

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
157

Fig. 4: A short circuit at a generating station

There are some other techniques adopting the concept of

the current limiter e.g. series inductor- thyristor-controlled [9],
the control regulates the average magnitude of the inductance
and hence its current limiting effect. One of the disadvantages
of this technique is the harmonic generation and the required
use of filters.
D.

Dividing the Bus Bar

Increasing the positive, negative, and zero sequence

impedances seen by the fault, will result in reducing the short
circuit currents. This is achieved in a substation by dividing
buses in two or more sections. This solution is quite common
and less expensive than upgrading circuit breakers. However,
load division requires a careful study since the operating
flexibility can be degraded in a significant manner. Also, the
reliability can be affected as the number of transformers
feeding the load is reduced.
In the system under study, this alternative was examined
only at the 132 kV stations to avoid losing the system
contingencies if applied on the BBs inside the generating
stations. Even with the 132 kV stations, the application is
limited to stations which have 4 275/132 kV transformers
where the load can be divided equally inside the station. These
conditions are not fulfilled in most of the stations.
E. etwork Splitting
If the power system enters the emergency mode due to
some disturbances, emergency controls will be applied to
drive the system back to the alert mode. If the emergency
controls fail to bring the system parameters back within their
constraints, protective devices will act to prevent the power
system components operating under unacceptable conditions

from undergoing damages [10, 11]. This may lead to system

islanding, which often produces unstable or generation- load
unbalanced islands. One quick solution is to temporarily split
the power system
The most difficult part in the study of system splitting
problem (SS problem) is to determine the proper splitting
points (called splitting strategies) to split the entire
interconnected transmission network into islands ensuring
generation/load balance and satisfaction of transmission
capacity constraints when islanding operation of the system is
unavoidable. For a large-scale power system, the SS problem
is very complicated in general. Moreover, network splitting
presents a temporary (not guaranteed) solution to the
developing SC problem.
IV. New Transmission Network Planning
The conclusions from the previous studies show that none
of these techniques is perfect in solving the developed
problem of increasing short circuit current levels. Taking into
consideration that there are at least two new generating
stations to be installed within few years, the optimum solution
for the Kuwait network in solving the problem of the
increasing SC levels is to construct a new higher voltage
network (400 kV) and to connect all the new generating
stations to this network. At the same time, the existing
275/132 kV network must be split permanently into three
networks connected to the new 400 kV as shown in Fig. 5. The
SCC, by this way, can be effectively reduced to acceptable
levels.
To reach the above stated target, various network
topologies are considered. The calculations performed include
alternatively applied SCC, load flow and contingency analysis

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
158

References
[1] J. Panek, H. Elahi, "Substation Voltage Upgrading",
IEEE Transactions on Power Delivery, Vol. 4, No. 3, July
1989, pp. 1715- 1724.

New ZNPS

GS

[2] ASPEN-OneLiner Simulation Program.

400/275 kV

The 3 spitted networks

FINT

New 400 kV Network

for the proposed configuration in order to prove the technical

requirements.

JABR

400/275 kV

SLBY

400/275 kV

[3] H. Rahman and B. H. Khan, "Power Upgrading of

Transmission Line by Combining ACDC Transmission",
IEEE Transactions on Power Systems, Vol. 22, No. 1,
February 2007, pp. 459- 466.
[4] Petersson, A., A. Edris, Dynamic Performance of the
Eagle Pass back-to-back HVDC Light Tie, 7th Intentional
conf on AC-DC Power Transmission, 2001, pp.220-225.
[5] Alex Y. Wu, Yuexin Yin, " Fault Current Limiter
Applications in Medium and High-Voltage Power Distribution
Systems", IEEE Transactions on Industry applications, Vol.
34, No. 1, January/February 1998, pp. 236-242.
[6] Mehrdad Tarafdar Haghl and Mehdi Abapour, "A New
Topology of Fault-current Limiter and its control strategy",
SICE-ICASE International Joint Conference, Oct. 18-21, 2006
, Bexco, Busan, Korea, pp. 5530-5534.

GS
New SWPS

Fig. 5: The proposed power system

V. Conclusions
High level of fault currents is a normal consequence of
transmission network expansion that follows the power system
development. High fault levels require switchgear and other
equipment with high rupture capacities. Reducing the fault
level in a power system is cheaper alternative compared to the
replacement of the switchgear. This paper presents different
measures examined in order to control the high SC current at
Kuwait high voltage network. Constructing a 400 kV high
voltage network with a permanent Splitting of the existing 275
kV network, prove to be the most suitable solution for the
highly increasing SC levels.

Acknowledgments
The financial support from the General Authority for
Applied Education and Training in Kuwait (Project No. TS02-08 ) is highly appreciated.

[7] S. K. Lowe, "Circuit Breaker Stresses Associated with

H.V. Series Reactors", Electrical Energy Conference,
Australian Electrical Research, Canberra 10-12 May 1978, pp.
114-119.
[8] K. Hongesombut, Y. Mitani, and K. Tsuji, "Optimal
Location Assignment and Design of Superconducting Fault
Current Limiters Applied to Loop Power Systems", IEEE
Transactions on Applied Superconductivity, Vol. 13, No. 2,
June, 2003 pp 1828-1831.
[9] Parihar, P. G.G. Karady, Characterization of a
thyristor controlled reactor, Electric Power Systems Research,
No.4, 1997, pp. 141-149.
[10] Ming Jin, Tarlochan S. Sidhu, and Kai Sun, " A New
System Splitting Scheme Based on the Unified Stability
Control Framework", IEEE Transactions on Power Systems,
Vol. 22, NO. 1, Feb. 2007, pp. 433-441.
[11] Kai Sun, Da-Zhong Zheng, and Qiang Lu, "Splitting
Strategies for Islanding Operation of Large-Scale Power
Systems Using OBDD-Based Methods", IEEE Trans. on
Power Systems, Vol. 18, No. 2, May 2003, pp. 912-923.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
159

IMPACT OF INVERTER-FED MULTI-MOTOR

DRIVES ON THE QUALITY OF ELECTRIC
POWER IN THE MAINS.
Maciej Pawowski

University of Technology Wroclaw,Faculty of Electrical Engineering

Plac Piastowski 27,58560 Jelenia Gra
[Link]@[Link], Tel/fax+(48)757551599

Abstract-The

article presents the results of analysis of higher

harmonics of current and voltage in networks feeding six-puls
[Link] dependences of this phenomena at inductance in a
dc circuit is also given. evaluation of this applicability of the analysis
for practical purposes, verification measurements is [Link] is done
measurements on the reel cutter of a paper-making machine mp-IV in
wiecie.

I.

INTRODUCTION.

An increased share of power of non-linear receivers

on the power network gives rise to significant transfiguration
of the feeding power sine waves. The performed analyses [2]
[3] clearly indicate that the degree of transfiguration of the
feeding power is higher at definite times of day. This refers to
user supply networks, on which the number of small recipients,
like PC users, significantly increases. Upper harmonics are
common in electric networks Their sources are all sorts of nonlinear receivers like converter systems, arc furnaces etc. The
receivers get a transmuted current and thus become sources of
upper harmonic currents [4],[5]. Those currents cause voltage
drops at network impedance and distort voltage sinusoid
[2],[6],[7]. However, it seems important to suggest a relevant
model of power supply network including linear and non-linear
receivers and a capacitor set. It is even more vital to discuss the
practicality of applying some proposed models. In converter
drives, the rectifier load current is an oscillating one since the
value of inductance in a DC circuit is limited. The phenomenon
has a significant impact on the harmonic spectrum of the input
current fed to the rectifier from the mains. Of course,
converters are usually equipped with commutation reactors, so
the commutation angle is not zero. Those factors are
especially important with some electric drives, as due to the
high circuit dynamics required, the drives have commutation
reactors but no smoothing reactors in the load circuits[8],[9].
II.

SPECIFICATIONS OF SUPPLY SYSTEMS FOR

INDUSTRIAL PLANTS

Mixed-type networks are quite frequently used, especially

when the internal network has suffered from problems with

compensating for wattless power or with the negative impact of

non-linear recipients on power supply network. The impact
involves taking transfigured power from the network, which
means that the power spectrum will include higher harmonics
[10], [2]. So far the analyses of upper harmonics of current
have usually neglected the resistance values of all supply
system components due to an assumption that their impact on
the value of upper harmonics is negligible. The size of
distortion of incoming voltage is determined by two major
factors, i.e.:
-type of non-linear source of current
-parameters of supply network
Fig. 1 illustrates a standard network supply system.

Tr1

Tr2

C
M

Tr3

Tr4

Fig. 1 Standard network supply system

In converter drives, the rectifier load current is

an oscillating one since the value of inductance in a DC circuit
is limited. The phenomenon has a significant impact on the
harmonic spectrum of the input current fed to the rectifier from
the mains. Of course, converters are usually equipped with
commutation reactors, so the commutation angle is not zero.
Those factors are especially important with some electric
drives, as due to the high circuit dynamics required, the drives

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
160

have commutation reactors but no smoothing reactors in the

load circuits[8],[9]. Fig.2 shows a schematic diagram of a sixpulse rectifier fed from a transformer of D vector group as
well as a current waveform iW ( x ) in the transformers
secondary at any phase.

compensate for passive power. Its module will be determined

as follows [3]:
1
1
1
1
1
1
=
+
+
+
+j
Zz
Z L Z S RO jX O
XC
1
(4)
where
Z L = RL + jX L - impedance of supply system,

Z S = RS + jX S - impedanc of linear recivers

III.

Fig.2. Schematic diagram of a six-pulse rectifier and a waveform of current at

any phase, where:
G generators
M DC induction motor
1, 2, 3 to 6 valves in individual bridge branches.

The oscillation of load current of a rectifier drive depends not

only on the load reactance X d , as it is frequently assumed, but
rather on a sum of the load circuit reactance and the
commutation reactance X d + 2 X K . If it is assumed that nonlinear receivers are 6-pulse SCR transformers (with high
shorting power SZ), then the feeding current will be:
1
1
1
ip (t ) = 2I1 (cost cos5t + cos 7t cos11t + ...)
5
7
11

MEASUREMENT-BASED ANALYSIS OF THE

HARMONICS PHENOMENON

In the process of evaluating the inter-harmonics amplitude

range, the harmonics in the power fed to an MP2 paper-making
machine operated in Skolwin were measured and analysed. The
machine power supply circuit beyond the marked measurement
points is shown in Fig. 3.
The measurement points were established at clamps feeding
PANT six-pulse rectifier 500 kW T7.T8 and on the clamps of
the secondary winding of Tr7a transformer feeding the group
of 1100kW total installed power. Some measurement points
were also located in the secondary winding of Tr7 transformer
feeding a group of inverters of 1000kW total installed power.
Measurements of the harmonics levels for loads and voltages
on the primary windings of those transformers.

(1)

where:

i p (t ) is a momentary amperage value at the pri of a

transformer, and I 1 is an effective value of the first harmonic.

In the system presented in Fig.1, the passage of non-linear
current In will cause voltage drops at the resultant impedance
of Zn system. Those voltage drops coincide with the supply
voltage sinusoid and cause its distortion. Therefore we can say
that

U n = I n Z n
where

Z n = Rn2 + n X L50

(2)

(3)

U n - voltage drop with a harmonic of n-th order at the

impedance of Zn system,

Z n - module of system impedance

with a harmonic of n-th order,

I n - module of power with a

harmonic of n-th order, Rn - system resistance with harmonics

X L50 - system inductive reactance with the basic

harmonic of 50Hz, n - harmonic order. The value of resultant

of n-th order,

impedance of Zn system will change when upper harmonics

filters or a capacitor set is incorporated into the system to

[Link] diagram of the power supply circuit in the A11 switching

station feeding the MP2 paper-making machine in Skolwin.

The measurements were performed on a system with an

open bus bar and the Tr7a transformer fed from the GII
generator at own power plant. Such a measurement setup
allows protection of a tested circuit from any impact of the
supply system. For comparison purposes, the Tr7 transformer
was supplied from the commercial power system.
IV. ANALYSIS OF MEASUREMENT RESULTS.
Measurement results are usually presented in the form of a
spectrum of typical sometimes non-typical harmonics that
are integral multiples of the base frequency [ 1 = 50 Hz ].
This results from the fact that the commonly used measurement
instruments use a phaselocked loop and sample the analysed

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
161

signal during one or several wave periods. The analysis is done

with the Fast Fourier Transform (FFT) or sometimes with the
Discrete Fourier Transform (DFT). With inter-harmonics
analysis the problem is more complicated since their
frequencies are not integral multiples of the base frequency and
moreover they often change during the measurements. They
manifest themselves as sidebands around the characteristic
harmonics. In the beginnings, their appearance would be
considered a measurement error and then as a methodological
error. Nowadays their is no doubt that those frequencies are
actually present in the system [12],[13].
Figure 4 presents current waveform in the T phase of Tr7a
transformer feeding the inverter set of MP2 machine drive at
the paper mill in Skolwin as well as its spectrum for varied
periods of measurement averaging.

Current waveform in the T phase

It is clearly visible in Fig.5 which shows the waveforms for

one of the component drives in the same switching station, this
time in the R phase,when the amplitude of upper harmonics of
current reach 45% of the base.
Similar results were obtained from measurements of other
components of a multi-motor drive, as shown in Fig.6
Interesting results were obtained from measurements of the
screen drive current, as shown in [Link] this particular case
the feeding system apparently works in near-resonance
conditions, as proved by the oscillations overlapping with the
current [Link] amplitudes reach 32% in this
case.

Current waveform in the S phase Current spectrum for the averaging period of
183.6ms.
[Link] and amplitude-frequency spectrum of the current feeding the
MP2 screen roll drive Switching station A11, transformer Tr7a, band 2kHz,
ASEA cabinet

Similar results were obtained from measurements on the

primary winding of the feeding transformer. Fig.7 presents the
current waveform on the 6kV side of the tested transformer and
its spectrum.
Current spectrum for the averaging period 40ms.

Current spectrum for the averaging period 115ms.

[Link] and amplitude-frequency spectrum of the current feeding the
MP2 machine drive Switching station A11, transformer Tr7a, band 2kHz,
cabinet no.1

There are well visible sidebands around all the frequencies that
are integral multiples of the base frequency. Their amplitudes
reach 1.5%. When comparing spectra for various averaging
periods, it becomes evident that the spectrum share of those
harmonics increases with the length of averaging periods.
Another factor affecting the size of inter-harmonics amplitude
is the level of distortions in the current fed from the mains.

Current waveform in the R phase.

Current waveform for the averaging period

of 40ms.
[Link] and amplitude-frequency spectrum of the current feeding
drives 1 and 2 of the MP2 press. Switching station A11, transformer Tr7a,
band 2kHz, cabinet no1

Current waveform in the S phase

Current waveform for the averaging
(6kV)
period of 145ms
[Link] and amplitude-frequency spectrum of the current feeding the
MP2 machine drive Switching station A11, transformer 7a, band 2kHz, cabinet
no.1, 6kV level

Conclusions from this measurement are basically identical

as those for the secondary winding, even though attenuation by
transformer reactance X Tr (k ) could be expected. Some
attenuation is visible in the voltage harmonics, as shown in
Fig.8.

[Link] and amplitude-frequency spectrum of the voltage feeding 0,4

kVthe MP2 machine drive Switching station A11, transformer Tr7a, band
2kHz, cabinet no.1

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
162

CONCLUSIONS

1.A significant increase in the number of non-linear receivers

results in a significant decrease in electric power quality,
especially in multi-motor drive systems.
2. Analytical methods for assessing this phenomena are
complicated and their practical application is difficult.
Therefore the method of direct measurements is preferable.
3. Multi-motor systems with inverter-fed drives require
unconditional application of filters for higher harmonics of the
feeding current.
VI

DISCUSSION

The article does not include a discussion of the methodology of

current harmonics measurements, i.e. the impact of applying
various waveform analysers / measurement windows /
averaging periods on the measurement results. Differences in
measurement results that arise from the above are usually
neglected, but they can reach as much as several dozen percent.
This problem requires a deep analysis.

REFERENCES

[1] Arrilaga J.,Bradley D.,Bodger P.: Power systems [Link] Wiley

1985
[2] Pawowski M., Analiza oddziaywania tyrystorowych napdw maszyn
papierniczych na sie wewntrzn papierni, Rozprawa doktorska,
Politechnika Wrocawska 1982 r.
[3] Miedziski B. Okraszewski Z. Szkka S. Szymaski A., Wysze
harmoniczne jako rdo zagroe transformatorw zasilajcych odbiory
wielkiej mocy, Mechanizacja i automatyzacja grnictwa 1999 vol. 11.
[4] Szachtarin B. Artiuszin C. Goubiew C. Rukawica K., Rezonans i chaos
w odnoj nieliniejnoj systemie Elektricziestwo vol. 2/2000
[5] Miedziski B. Okraszewski Z. Szkka S. Szymaski A. Problemy
eksploatacyjne zwizane z wyszymi harmonicznymi prdu i napicia
generowane napdami duej mocy, Wgiel brunatny, wydanie specjalne
ISSN , pp. 1232 8782
[6] Jasiski J. Bartosiewicz M., Filtr aktywny do kompensacji mocy biernej
odksztacenia ,Materiay konferencyjne JEL 2000 r.
[7] Staszewski K., Wysze harmoniczne w sieci kopalnianej 6kV,
Mechanizacja i automatyzacja grnictwa vol. 7/367/2001
[8] Pirg S., Energoelektronika ,AGH Uczelniane Wydawnictwa Naukowo
- [Link] 1998.
[9] MoltgenG., Netzgefuerte Stromrichter mit Thyristoren,Siemens
[Link] 1974
[10] Nowak M. Barlik R. , Technika tyrystorowa , WNT Warszawa 1998 r.
[11] Pawowski M. Fjakowski Z., Rezonans prdw w sieciach
przemysowych 6kV, Mechanizacja i automatyzacja grnictwa
7/367/2001
[12] Maswood Ali J. ZHU JUN., Analysis of Harmonic motor controls
under nonlinear conditions , Electronics Machines and Power system
28/2000
[13] Maamar Bettaych, Urasi Qidwai., A new flux based harmonic
elimination technique, Electric Machines and Power System 27/1999
[14] Pawowski M. Szymaski A. Fjakowski Z. Zjawiska rezonansowe w
sieciach przemysowych Mechanizacja i automatyzacja grnictwa
2002 vol. 8.
[15] eelenko J.:Wysze harmoniczne w systemach zasilajcych odbiorniki
Przemysowe [Link] 1994

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
163

Modeling analysis and solution of Power

Quality Problems
Mahesh Singh (Lecturer) SSCET BHILAI INDIA
Vaibhav Tiwari CSIT BHILAI INDIA

Abstract-- A Power quality problem is an occurrence

manifested as a nonstandard voltage, current or frequency that
results in a failure or a mis-operation of end use equipments.
Utility distribution networks, sensitive industrial loads, and
critical commercial operations all suffer from various types of
outages and service interruptions which can cost significant
financial 1oss per incident based on process dow n-time, lost
production, idle work forces, and other factors. With the
restructuring of Power Systems and with shifting trend
towards Distributed and Dispersed Generation, the issue of
Power Quality is going to take newer dimensions. The aim
therefore, in this work, is to identify the prominent concerns in
the area and thereby to recommend measures that can enhance
the quality of the power, keeping in mind their economic
viability and technical repercussions. In this paper
electromagnetic transient studies are presented for the
following two custom power controllers: the distribution static
compensator (D-STATCOM), and the dynamic voltage
restorer (DVR). Comprehensive results are presented to assess
the performance of each device as a potential custom power
solution.
Index Terms-- Power Quality Problems, Power System
Restructuring, Voltage Sag, DSTATCOM, DVR, MATLAB.

I. INTRODUCTION
ower quality is certainly a major concern in the present
era; it becomes especially important with the
introduction of sophistic ated devices, whose
performance is very sensitive to the quality of power supply.
Modern industrial processes are based a large amount of
electronic devices such as programmable logic controllers
and adjustable speed drives. The electronic devices are very
sensitive to disturbances [1] and thus industrial loads
become less tolerant to power quality problems such as
voltage dips, voltage swells, and harmonics.

Voltage dips are considered one of the most severe

disturbances to the industrial equipmen t. A paper machine
can be affected by disturbances of 10% voltage drop lasting
for 100ms. A voltage dip of 75% (of the nominal voltage)
with duration shorter than 100ms can result in material loss
in the range of thousands of US dollars for the

semiconductors industry [2]. Swells and over voltages can

cause over heating tripping or even destruction of industrial
equipment such as motor drives. Electronic equipments are
very sensitive loads against harmonics because their control
depends on either the peak v alue or the zero crossing of the
supplied voltage, which are all influenced by the harmonic
distortion.
This paper analyzes the key issues in the Power
Quality problems, specially keeping in mind the present
trend towards more localized generations (also termed as
distributed and dispersed generation) and consequent
restructuring of power transmission and distribution
networks. As one of the prominent power quality problems,
the origin, consequences and mitigation techniques of
voltage sag problem has been discussed in detail. The study
describes the techniques of correcting the supply voltage sag
in a distribution system by two power electronics based
devices called Dynamic Voltage Restorer (DVR) and
Distribution STATCOM (D-STATCOM).A DVR voltage in
series with the system voltage and a D -STATCOM injects a
current into the system to correct the voltage sag [1]. The
steady state performance of both DVR and D-STATCOM is
studied for various levels of voltage sag levels.
II. SOURCES AND EFFECTS OF POWER QUALITY
PROBLEMS
The distortion in the quality of supply power can be
introduced /enhanced at various stages; however, some of
the primary sources of distortion [3] can be identified as
below:
A.
B.
C.
D.
E.
F.
G.
H.

Power Electronic Devices

IT and Office Equipments
Arcing Devices
Load Switching
Large Motor Starting
Embedded Generation
Electromagnetic Radiations and Cables
Storm and Environment Related Causes etc.

While power disturbances occur on all electrical

systems, the sensitivity of todays sophi sticated electronic
devices makes them more susceptible to the quality of
power supply. For some sensitive devices, a momentary
8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
disturbance can cause scrambled data, interrupted
164

communications, a frozen mouse, system crashes and

equipment failure etc. A po wer voltage spike can damage
valuable components. Some of the common power quality
issues and their prominent impact are summarized in the
table below:

needed. It has a dynamic response time of about one

cycle. To correct quickly for voltage spikes, sags or
interruptions, the static switch can used to switch one or
more of devices such as capacitor, filter, alternate power
line, energy storage systems etc. The static switch can be
used in the alternate power line applications. T his scheme
requires two independent power lines from the utility or
could be from utility and localized power generation like
those in case of distributed generating systems [4]. Such a
scheme can protect up to about 85 % of interruptions and
voltage sags.
C. Energy Storage Systems:
Storage systems can be used to protect sensitive
production equipments from shutdowns caused by voltage
sags or momentary interruptions. These are usually DC
storage systems such as UPS, batteries, superconducting
magnet energy storage (SMES), storage capacitors or
even fly wheels driving DC generators [6]. The output of
these devices can be supplied to the system through an
inverter on a momentary basis by a fast acting electronic
switch. Enough energy is fed to the system to compe nsate
for the energy that would be lost by the voltage sag or
interruption. In case of utility supply backed by a
localized generation this can be even better accomplished.

Table1: Various power quality problems and their effects

III. SOLUTIONS TO POWER QUALITY PROBLEMS

There are two approaches to the mitigation of power
quality problems. The solution to the power quality can be
done from customer side or from utility side [4]. First
approach is called load conditioning, which ensures that
the equipment is less sensitive to powe r disturbances,
allowing the operation even under significant voltage
distortion. The other solution is to install line conditioning
systems that suppress or counteracts the power system
disturbances. A flexible and versatile solution to voltage
quality problems is offered by active power filters.
Currently they are based on PWM converters and connect
to low and medium voltage distribution system in shunt or
in series. Series active power filters must operate in
conjunction with shunt passive filters in ord er to
compensate load current harmonics. Shunt active power
filters operate as a controllable current source and series
active power filters operates as a controllable voltage
source. Both schemes are implemented preferable with
voltage source PWM inverter s [5], with a dc bus having a
reactive element such as a capacitor. Active power filters
can perform one or more of the functions required to
compensate power systems and improving power quality.
Their performance also depends on the power rating and
the speed of response.
However, with the restructuring of power sector and with
shifting trend towards distributed and dispersed
generation, the line conditioning systems or utility side
solutions will play a major role in improving the inherent
supply quality; some of the effective and economic
measures can be identified as following:

D. Electronic tap changing transformer:

A voltage-regulating transformer with an elec tronic load
tap changer can be used with a single line from the utility.
It can regulate the voltage drops up to 50% and requires a
stiff system (short circuit power to load ratio of 10:1 or
better). It can have the provision of coarse or smooth steps
intended for occasional voltage variations.
E. Harmonic Filters
Filters are used in some instances to effectively reduce
or eliminate certain harmonics [7]. If possible, it is always
preferable to use a 12 -pluse or higher transformer
connection, rather than a fil ter. Tuned harmonic filters
should be used with caution and avoided when possible.
Usually, multiple filters are needed, each tuned to a
separate harmonic. Each filter causes a parallel resonance
as well as a series resonance, and each filter slightly
changes the resonances of other filters.
F. Constant-Voltage Transformers:
For many power quality studies, it is possible to greatly
improve the sag and momentary interruption tolerance of
a facility by protecting control circuits. Constant voltage
transformer (CVTs) can be used [6] on control circuits to
provide constant voltage with three cycle ride through, or
relays and ac contactors can be provided with electronic
coil hold-in devices to prevent mis-operation from either
low or interrupted voltage.

A. Lightening and Surge Arresters:

Arresters are designed for lightening protection of
transformers, but are not sufficiently voltage limiting for
protecting sensitive electronic con trol circuits from
voltage surges.

G. Digital-Electronic and Intelligent Controllers for

Load-Frequency Control:
B. Thyristor Based Static Switches:
Frequency of the supply power is one of the major
The static switch is a versatile device for switching a
determinants
of power quality, which affects the
8. EEEIC
International
Conference
on Environment
and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
new element
into the
circuit when
the voltage
support is
165

equipment performance very drastically. Even the major

system components such as Turbine li fe and
interconnected-grid control are directly affected by power
frequency. Load frequency controller used specifically for
governing power frequency under varying loads must be
fast enough to make adjustments against any deviation. In
countries like India and other countries of developing
world, still use the controllers which are based either or
mechanical or electrical devices with inherent dead time
and delays and at times also suffer from ageing and
associated effects. In future perspective, such cont rollers
can be replaced by their Digital -electronic counterparts.
IV. USE OF CUSTOM POWER DEVICES TO IMPROVE
POWER QUALITY
In order to overcome the problems such as the ones
mentioned above, the concept of custom power devices is
introduced recently; custom power is a strategy, which is
designed primarily to meet the requirements of industrial
and commercial customer. The concept of custom power
is to use power electronic or static controllers in the
medium voltage distribution system aiming to supply
reliable and high quality power to sensitive users [1].
Power electronic valves are the basis of those custom
power devices such as the static transfer switch, active
filters and converter-based devices. Converter based
power electronics devices can be divided in to two groups:
shunt-connected and series-connected devices. The shunt
connected devices is known as the DSTATCOM and the
series device is known as the Static Series Compensator
(SSC), commercially known as DVR. It has also been
reported in literature that both the SSC and DSTATCOM
have been used to mitigate the majority the power system
disturbances such as voltage dips, sags, flicker unbalance
and harmonics.
For lower voltage sags, the load voltage magnitude can
be corrected by injecting only reactive po wer into the
system. However, for higher voltage sags, injection of
active power, in addition to reactive power, is essential to
correct the voltage magnitude [8]. Both DVR and
DSTATCOM are capable of generating or absorbing
reactive power but the active power injection of the device
must be provided by an external energy source or energy
storage system. The response time of both DVR and D STATCOM is very short and is limited by the power
electronics devices. The expected response time is about
25 ms, and which is much less than some of the
traditional methods of voltage correction such as tap changing transformers.

will be taken up for analysis and performance comparison

of these devices. The modeling approach adopted in the
paper is graphical in nature, as opposed to mathematical
models embedded in code using a high-level computer
language. The well-developed graphic facilities available in
an industry standard power system package, namely,
MATLAB (/Simulink) [12], is used to conduct all aspects
of model implementation and to carry out extensive
simulation studies.
The control scheme for these devices is shown in Fig.1.
The controller input is an error signal obtained from the
reference voltage and the value rms of the terminal voltage
measured. Such error is processed by a PI controller and
the output is the angle , which is provided to the PWM
signal generator. The PWM generator then generates the
pulse signals to the IGBT gates of voltage source converter
[10].

Fig.1. The PI Controller

(A) D-STATCOM
The test system employed to carry out the simulations
concerning the DSTATCOM actuation for voltage s ag
compensation is shown in Fig.2. Such system is composed by
a 230 kV, 50 Hz transmission system, represented by a
Thevenin equivalent, feeding a distribution network through a
3-winding transformer connected in Y/Y/Y, 230/11/11 kV.
To verify the working of a DSTATC OM, a variable load is
connected at bus 2. During the simulation, in the period from
500 to 900 ms, the switch S1 is closed .The above test system
is simulated under the environment of Matlab -Simulink and
power system block set (PSB) the model used for this purpose
is shown in the fig.3.

V. MODELING OF CUSTOM POWER DEVICES AND

SIMULATION RESULTS
As mentioned in the previous section that custom power
devices could be the effective means to overcome some of
Fig.2. Test system for DSTATCOM
the major power quality problems by the way of injecting
active and/or reactive power(s) into the system [9]-[11].
This section of the paper deals with the modeling of
8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
DSTATCOM
and DVR. Consequently some case studies166

Fig.3. MATLAB Simulation model of DSTATCOM

A set of simulations was carried out for the test system

shown in Fig.2. The simulations relate to three main operating
conditions.
1) In the simulation period 500 900 ms, the load is increased
by closing switch S1. In this case, the voltage drops by almost
27% with respect to the reference value.
2) At 900 ms, the switch S1 is opened and remains so
throughout the rest of the simulat ion. The load voltage is very
close to the reference value, i.e., 1 pu.
3) In order to gain insight into the influence that capacitor
size has on D-STATCOM performance, simulations were
carried out with different size of capacitors . The total
simulation period is 1.4 s.

When the Switch S1 closes, the D-STATCOM supplies

reactive power to the system. By way of example, Fig.5(a).
shows the regulated rms voltage corresponding to a 750 F
capacitor, where a rapid regulation response is obtained and
transient overshoots are almost nonexistent.

Fig 5(a) Voltage response of the test system 1 with DSTATCOM;

With 750F Capacitor

Fig.4. Voltage response of the test system without DSTATCOM

Fig.4. shows the rms voltage at the load point for the case
when the system operates with no D -STATCOM. Similarly, a
new set of simulations was carried out but now with the D STATCOM connected to the system. The results are shown in
Fig.5(a). Where the very effective voltage regulation provided
by the D-STATCOM can be clearly appreciated.

Fig 5(b) Voltage response of the test system 1 with DSTATCOM;

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz,
May 10-13, 2009
With Poland,
75F Capacitor
167

This contrasts with cases where the capacitor is undersized.

For instance, Fig. 5(b) shows the rms voltage for the case
when a 75 F capacitor is employed.
(B) DVR
The test system employed to carry out the simulations
concerning the DVR actuation is shown in Fig. 6. Such
network is composed by a 13 kV, 50 Hz generat ion system,
represented by a Thevenin equivalent, feeding two
transmission lines through an 3-winding transformer
connected in Y// 13/115/115 kV. Such transmission lines
feed two distribution networks through two transformers
connected in /Y, 115/11 kV. To verify the working of a
DVR employed to avoid voltage sags during short -circuit, a
fault is applied at point X via a resistance of 0.4 . Such fault
is applied from 500 to 900 ms.

Fig.6. Test system for DVR

The above test system is simulated under the environment

of Matlab-Simulink and power system block set (PSB) the
model used for this purpose is shown in the fi g.7.

Fig.7 MATLAB Simulation model of DVR

Two simulations studies are carried out as follows:

1) The first simulation contains no DVR and a three -phase
short-circuit fault is applied at point A, via a fault resistance
of 0.4 , during the period 500 900 ms. The voltage sag at the
load point is almost 60% with respect to the reference
voltage.
2) The second simulation is carried out using the same
scenario as above but now with the DVR in operation. The
total simulation period is 1.4 s.
Using the facilities available in MATLAB the DVR is
simulated to be in operation only for the duration of the fault,
as it is expected to be the case in a practical situation. The
results for both simulations are shown in Fig .8 and Fig.9
Voltage response without DVR is shown in Fig.8

Fig.8 Voltage Response of the test system without DVR.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
168

When the DVR is in operation the voltage sag is mitigated

almost completely, and the rms voltage at the sensitive load
point is maintained at 98%, as shown in Fig. 9.

VII. REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]

Fig 9 Voltage Response of the test system2 with DVR.

The PWM control scheme controls the magnitude and the

phase of the injected voltages, restoring the rms voltage very
effectively [13]-[14]. The sag mitigation is performed with a
smooth, stable, and rapid DVR response;
VI. CONCLUSIONS
Power quality measures can be applied both at the user end
and also at the utility level. The work identifies some
important measures that can be applied at the utility level
without much system upset (or design changes).This pape r
has presented models of custom power equipment, namely D STATCOM, DVR, and applied them to mitigate voltage dip
which is very prominent as per utilities are concerned. The
highly developed graphic facilities available in MATLAB
SIMULINK were used to cond uct all aspects of model
implementation and to carry out extensive simulation studies
on test systems. A new PWM -based control scheme has been
implemented to control the electronic valves in the two -level
VSC used in the D-STATCOM and DVR. As opposed to
fundamental frequency switching schemes already available
in the MATLAB SIMULINK. This characteristic makes it
ideally suitable for low-voltage custom power applications. It
was observed that in case of DSTATCOM capacity for power
compensation and voltage re gulation depends mainly on the
rating of the dc storage device. The simulation results
presented shows good accuracy with results reported in index
journals.

[9]
[10]
[11]
[12]
[13]
[14]

H. Hingorani Introducing custom power IEEE spectrum, vol.32 no.6

June 1995 p 41-48.
Ray Arnold Solutions to Power Quality Problems power engineering
journal 2001 pages: 65-73.
John Stones and Alan Collinsion Introduction to Power Quality
power engineering journal 2001 pages: 58 -64.
Gregory F. Reed, Masatoshi Takeda, "Improved power quality
solutions using advanced solid -state switching and static compensation
technologies," Power Engineering Society 1999 Winter Meeting, IEEE
G. Venkataramanan and B. Johnson, A pulse width modulated power
line conditioner for sensitive load centers, IEEE Trans. Power
Delivery, vol. 12, pp. 844849, Apr. 1997.
N.G. Hingorani and L. Gyugyi, Understanding FACTS: Concepts and
Technology of Flexible AC Transmission Systems, 1st edition, The
Institute of Electrical and Electronics Enginee rs,2000.
F. Z. Peng, H. Akagi, and A. Nabae, Compensation characteristics of
the combined system of shunt passive and series active filters, IEEE
Trans. Ind. Applicat., vol. 29, pp. 144151, Jan./Feb. 1993.
[Link] Compensation of distribution system voltage sag by DVR
and DSTATCOM Power Tech Proceedings, 2001 IEEE Porto , Volume:
1 , 10-13 Sept. 2001 Pages:5 pp. vol.1
S. S. Choi, B. H. Li, and D. [Link], Dynamic voltage
restoration with minimum energy injection, IEEE Trans. Power Syst.,
vol. 15,pp. 5157, Feb. 2000
O. Anaya-Lara, E. Acha, Modeling and Analysis of Custom Power
Systems by PSCAD/EMTDC, IEEE Trans., Power Delivery, PWDR
vol-17 (1), pp. 266-272, 2002.
P.W. Lehn and MR. Iravani, Experimental Evaluation of STATCOM
Closed Loop Dynamics, IEEE Trans. Power Delivery, vol. 13, no.4,
October 1998, pp.1378-1384.
TEQSIM International Inc., Power System Blockset User'sGuide,
2001.
M. H. J. Bollen, Understanding Power Quality Problems Voltage
Sags and Interruptions Piscataway, New York: IEEE Press, 2000.
Mohan, T. M. Undeland, andW. P. Robbins, Power Electronics:
Converters,Applications and Design . New York: Wiley, 1995.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
169

Power System Real Time Simulation Testing Using IEC 61850.

* Yogesh Sonawane Pankaj Patil
yogimba@[Link]

pankajpatil2k@[Link]

[Link]

Sayali Sonawane

starnaresh2k@[Link] sayali001@[Link]

* Corresponding Author.

Abstract:

IEC-61850

becomes more widely accepted in the

electrical engineering community, it is important that
the testing tools keep pace with this development. IEC
61850 presents new challenges to real time simulation
and closed-loop testing of protective relays. Electrical
interfaces used for binary signaling and voltage/current
amplifiers must be replaced by an Ethernet connection
and an IEC 61850 protocol stack. The electrical
interfaces of a real time simulator are engineered to
provide low latency and deterministic performance
appropriate for a real time simulation. Similar attention
must be given to IEC 61850 interfaces. Latency must be
minimized so that the IEC 61850 interface does not add
unacceptable delays to the operation of the simulator.
Also, protocol processing must be deterministic to allow
real time simulations to be repeatable and dependable.
In addition, IEC 61850 specifies new configuration
parameters and a new method for configuration called
the Substation Configuration Language (SCL). These
must be implemented in such a way that they fit within
the typical modes of operation of the simulator. The
paper presents a successful hardware implementation
for IEC 61850 messaging on a real time simulator and
discusses the key design criteria. The software required
to configure the IEC 61850 will also be addressed along
with the advantages in using the IEC 61850 protocol.
One of the biggest advantages is brought about by the
realization of the IEC 61850-9-2 sampled values,
removing the need for amplifiers as the standard
interface to protection devices. Sampled values of the
voltage and current signals can be sent via Ethernet,
making it even more practical to perform testing on a
protective relaying scheme rather than just individual
devices.
Keywords: IEC 61850, GOOSE, GSSE, closed-loop
testing, real time, power system simulation.

I. INTRODUCTION
Real time digital simulation using systems such
as the RTDS Simulator are often used to perform
closed-loop testing of power system protection devices
such as relays and protection schemes. IEC 61850 is
the international standard for substation automation
systems. It defines the communications between
devices in the substation and the related system
requirements. IEC 61850 supports all substation
automation functions and their engineering [1]. As
IEC 61850 becomes more widely accepted and
implemented by protection device manufacturers and
power system utilities, it is important that the testing

tools keep pace allowing these new devices to be

tested. Using IEC 61850 for simulation testing also
provides a number of benefits and advantages over
traditional electrical interfaces such as the
simplification of wiring and the potential to reduce the
cost of the test setup.
When using IEC 61850, the electrical interfaces
used for binary signaling and voltage/current
amplifiers are replaced by an Ethernet connection and
software implementing IEC 61850. Careful
consideration must be given to the integration of IEC
61850 into real time digital simulators to ensure that
the integrity of the simulation is not compromised due
to nondeterministic or slow performance of the IEC
61850 interface. This paper introduces real time digital
simulation testing that includes the use of IEC 61850.
A list of advantages for using IEC 61850 in simulation
testing is presented along with key design criteria used
when incorporating IEC 61850 into a real time digital
simulator.
II. SIMULATION TESTING USING REAL TIME
DIGITAL SIMULATORS
A. Common Applications
Real time digital simulators are commonly used
for closed loop testing of protection and control
devices as well as for real time system studies. When
conducting closed-loop testing, the simulator acts as
the power system and interfaces to the test objects. For
closed-loop operation protective relay testing, the
simulator must provide real time data (i.e. voltage,
current and breaker status) to the relay and sense trip
and reclose status from the relay. Since the power
system is being simulated, various faults can easily be
applied under different network conditions to evaluate
the performance of the protection and control. If the
protection detects the fault applied, the trip signal will
be sensed and the breaker in the simulation opened.
When evaluating the performance of a protective
relay or protection system, exhaustive testing is often
conducted. This involves applying thousands of faults
while varying different system parameters (e.g. fault
type, fault location, point-onwave). Since so many
tests are run, it is crucial to keep each test and the time
between tests as short as possible. Prior to the advent
of IEC 61850, the protection equipment was connected
to the simulator using individual wires. The
complexity of the wiring is evident even for the

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
170

relatively simple double ended line protection test

setup shown in Fig. 1. Twelve (12) signals have to be
connected from the RTDS.
Simulator to the voltage and current amplifiers.
After the signals pass through the amplifiers, twelve
(12) more connections have to be made to connect the
voltage and current signals to the relays. If the relays
are set for single pole trip and reclose, a minimum of
twelve (12) contacts must also be connected to the
simulators digital input to control breaker operation.
If internal relay elements are to be monitored, even
more contacts have to be connected to the simulator.
Finally, six (6) breaker status signals must be supplied
to the relays at station level voltages (e.g. 125 Vdc) via
potential free contacts. All told, there are 30 digital
and analogue connections between the relays and the
simulator when using traditional methods for the test
setup illustrated in Fig. 1. This does not include the 12
signals that must be connected from the simulator to
the amplifiers or the monitoring of internal relay
elements.

Fig 1. Test Setup with Standard Electrical Connections

to end transfer time is typically specified between 3-10

msec [3].
IEC 61850-9-2 Sampled Values can be used for
applying voltage and current signals to devices instead
of using analogue current and voltage from amplifiers.
Maximum sample to transmission time for an IEC
61850-9-2 sampled value message used for protection
is typically specified at less than 3 msec for sample
rates faster than 480 samples/cycle [3]. An IEC 61850
device referred to as a Merging Unit (MU) is used to
transmit sampled values derived from transducers
using IEC 61850-9-2 messages over the process bus.
Messages sent by MUs may contain an arbitrary
number of current and voltage channels/phases, but
initial implementations of 61850-9-2 send messages
with instantaneous 3-phase plus neutral voltage and
current samples (i.e. 8 signals) [2]. Arbitrary sample
rates are permitted by IEC 61850-9-2, but initial
implementations of MUs used for protection typically
send samples at 80 samples/cycle [3].
With the implementation of IEC 61850
communication for the RTDS Simulator, all signals
are passed to and from the protective relays via a
separate station bus and process bus both based on
Ethernet. The IEC 61850 GOOSE/GSSE facility
described in this paper allows up to 32 binary inputs
and 32 binary outputs to be provided to as many as 8
different IEDs (e.g. relays) from one device.
Therefore the trip and reclose signals plus the breaker
status signals for the double ended line protection tests
can be provided by one Ethernet connection
representing a single station bus. An added Ethernet
connection is required for a process bus to pass up to 8
sampled value (analogue) signals to each relay. A
double ended line protection test setup using IEC
61850 is shown in Fig 2. In the figure, the GTNET-SV
is used for the IEC 61850-9-2 interfaces and the
GTNET-GSE provides the GOOSE/GSSE interface.

B. Integration Of IEC 61850 In Simulation Testing

IEC 61850 can be used to replace the two types of
interfaces between the RTDS Simulator and the
devices being tested. IEC 61850 GOOSE/GSSE can
replace the hard-wired binary inputs and outputs over
what IEC 61850 refers to as a station bus and IEC
61850-9-2 sampled values can replace the hard-wired
voltage and current signals from the power amplifiers
over what IEC 61850 refers to as a process bus.
IEC 61850 GOOSE/GSSE replaces the hard-wired
binary signaling with a message sent over an Ethernet
based stations bus. All of the hard-wiring needed for
status signals can be replaced by a single Ethernet
connection between the simulator and the station bus
LAN to which the protection devices are connected.
The transfer of trip and breaker status through
Fig. 2. Test Setup with IEC 61850 Connections
8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
GOOSE/GSSE messages is very fast. Maximum end 171

III. ADVANTAGES
There are a number of advantages to using an
IEC 61850 interface for simulation testing compared
to using traditional electrical interfaces. An obvious
advantage is the ability to test IEC 61850 based
systems that cannot be tested using traditional
electrically interfaced simulators. The use of IEC
61850 in installations is increasing [4], and it is
important to be able to test these devices using the IEC
61850 interface. Also, since a goal of IEC 61850 is to
provide a comprehensive set of functions [3] to replace
proprietary protocols, simulation testing can be
performed on products from different vendors using
only the IEC 61850 protocol without requiring a
combination of industry standard and proprietary
protocols. Elimination of the often large and expensive
voltage and current amplifiers commonly used in
simulation testing reduces the size and cost of the test
setup. This in turn makes it more practical to perform
testing on larger protective relay schemes rather than
on individual devices.
Another major advantage of using IEC 61850
for simulation testing is the reduction in electrical
wiring between the simulator and protection devices.
Fig. 2 illustrates that when using IEC 61850 for the
same test system as shown in Fig. 1, the number of
connections between the simulator and the relays is
reduced to just 4. In addition, the connections are
greatly simplified by using an Ethernet connection
rather than individual wires and by eliminating the
need for station level breaker status signals.

The start-up operation of a simulation test case

also differs from other substation equipment.
Whenever possible, the simulation should require only
minimal time to start-up and initialize. This reduces
the amount of time to run a single simulation test case
and allows multiple simulation runs to be made in the
shortest period of time. In addition, the start-up and
initialization time should be as consistent as possible.
This is because the worst-case initialization delay must
be out-waited in order to ensure consistent
simulation results. The configuration of the simulator
is also different than most other substation equipment.
The simulator is frequently reprogrammed to perform
different simulations. In addition, the simulator is
frequently started, stopped and restarted. In
comparison, substation equipment is often set up
during commissioning and runs continuously for many
months without changes to the configuration other
than automatic changes to the settings of the protection
functions.

A. Hardware
The IEC 61850 interface for the RTDS
Simulator was implemented using a dedicated
processing platform. The only responsibility for this
processing platform is to handle IEC 61850
communications. Using a dedicated processing
platform for IEC 61850 communications allows
maximum hardware and software optimization to be
performed and allows control over input and output
latency both the amount and the variability from
packet to packet. This is because there is no sharing of
processing or communications resources between the
IV. IMPLEMENTATION OF AN IEC 61850
protocol stack and other unrelated simulation activity
INTERFACE FOR SIMULATION TESTING
Implementation of an IEC 61850 interface for a
as would be the case if the protocol processing was
power system simulator presents interesting challenges
integrated into hardware that was utilized for other
that are generally different than those present when
simulation functions.
implementing IEC 61850 in typical substation devices
When using a dedicated processing platform,
such as protection relays and circuit breakers.
overhead due to communications between the platform
A power system simulator must produce
and the rest of the simulator must be fast and efficient
reliable, consistent results. The results must be
so that total input and output packet latency is
consistent each time the same simulation case is run.
minimal. This can be achieved by utilizing a high
There must be low variability in the input and output
performance FPGA for the interface between the
latency and the latencies must be well understood to
protocol hardware and the simulator. Buffering and
allow the analysis of simulation results. The input and
filtering are easily implemented in an FPGA and
output latency must also be minimized to reduce the
provides an efficient interface for the processor and
impact of the interface and allow the widest range of
allows more CPU time to be spent processing packets.
delay adjustment within the simulation, if desired.
The accurate time-stamping required for IEC 61850-9Although latency must be minimized, it should not be
2 is realized by distributing a 1 pulse per second (1
minimized at the expense of variability in latency. Any
PPS) to the communicating devices. Implementing the
variation in latency affects the confidence in the
1 PPS input time synchronization functions within an
timing measurements of DUT operations, which in
FPGA eliminates processor overhead and dependency
8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
turn affect the
confidence in the simulation as a whole. 172 on processor interrupt latency. This allows the

processor to spend more time processing packets

rather than synchronizing to the 1 PPS input. A key to
minimizing the variability in input and output packet
latency is to reduce the amount of extraneous,
unexpected processing. The LAN can be a source of
significant unnecessary processing, but can be
minimized in most circumstances. The selection of the
Ethernet media access controller (Ethernet MAC) is
crucial in minimizing the impact of processing
unrelated network activity. The Ethernet MAC must
be equipped with the capability to perform packet
filtering based on a set of configured addresses.
Without hardware-based address filtering, each packet
on the network must be read by the CPU to determine
if the destination address in the message matches the
filter criteria. With hardware address filtering, only
packets with selected destination addresses are passed
to the processor. Although IEC 61850 does not use
Ethernet broadcast packets, such packets can be
present on networks and might affect overall network
performance. Disabling broadcast traffic during
simulation runs can provide further optimization. The
combination of hardware address filtering and
disabling broadcast traffic reduces the network traffic
entering to only those packets that are directly
addressed or requested there is no extra processing
even when the network is very active.

bus transmission of sampled values. Implementing

multiple types of messaging at the same time can
cause slight unexpected latencies. Since variability of
latency must be minimized in a real time simulator,

Fig. 3. GOOSE Reception analysis using RTOS Development Tool

each IEC 61850 interface device only runs one

type of messaging at a time. For example, a single IEC
61850 interface device can run GOOSE/GSSE or
61850-9-2 sampled values, but not both at the same
time. No interface related station level messaging is
supported while running GOOSE/GSSE or 61850-9-2
sampled values to avoid any possibility of affecting
latency. If required, multiple IEC 61850 interface
devices can be simultaneously connected to the
simulator to provide the different types of messaging,
while still providing excellent protocol performance.

B. Software
The RTDS Simulator implements IEC 61850
GOOSE, GSSE and 61850-9-2 Sampled Values under
C. Configuration
an off-the-shelf real time operating system (RTOS).
Digital simulators are commonly programmed
By utilizing a RTOS with a rich suite of development
for an application every time a simulation case is
tools, performance can be easily assessed and
downloaded. For this reason, all details for the IEC
optimized. Fig. 3 shows a capture by one of the RTOS
61850 interface must be determined prior to
tools and shows the processing of a single received
downloading a test case. When using simulators in
GOOSE message. From this capture the approximate
batch mode, it must be possible to easily change all
processing times, the number of context switches, and
simulation parameters from a script file. IEC 61850
all operating system calls such as semaphore and
configuration of the simulator is easily accomplished
message queue activity can be determined. This type
using the IEC 61850 SCL features described in IEC
of capture can be used to analyze and optimize the
61850-6 [3].
processing of packets.
SCL is an extensible markup language (XML)
A third-party IEC 61850 protocol stack was
based description language used to describe IEC
chosen to allow development time to be spent on
61850 IED configurations and communications. A
integration and optimization rather than the
single SCL file can contain a configuration and
implementation of intimate details of the protocol. Full
description of each IED so that the overall simulation
source code was provided with the protocol software
configuration can be stored in a single SCL file. An
so that it was possible to verify the protocol processing
IEC 61850 system engineering software package is
was implemented efficiently. The IEC 61850 protocol
used to modify the system configuration and generate
covers a wide range of functionality from relatively
an SCL file to be read by the configuration software
interface related station level reporting of system
for each connected IED. The system configuration
values and status to the very fast process related
function can be performed by many different system
EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
station [Link]/GSSE
messaging and the process 173 engineering tools since SCL is not proprietary. The

compiler for the RTDS Simulator reads SCL files to

determine the configuration and connections for the
IEC 61850 interface. An SCL file can be generated for
each test scenario and can be selected through
simulator scripts. This allows compatibility between
IEC 61850 and the simulator batch mode operation.
Some IEC 61850 stacks provide advanced features
such as the automatic discovery of multicast
destination addresses used by remote IEDs to simplify
configuration. These features should be avoided for
simulation testing if they affect the deterministic
operation of the IEC 61850 interface or if they
increase initialization time at the beginning of
simulation. Automatic discovery of multicast
addresses requires the IEC 61850 interface to process
all multicast traffic on the network until a multicast
message is received from the designated remote IED.
This increase in processing can cause undesirable
variations in input and output latency, and must be
avoided in a simulation testing environment

Standards:
[3] IEC 61850 Communication networks and systems in
substations All parts, Reference number IEC 61850-SER.
Papers Presented at Conferences:
[4] C. Hoga, G. Wong, "Utilities and Industries of Today: Leading
by Following IEC 61850," presented at the 15th Power Systems
Computation Conference, Liege, 2005.
[5]Jan.23-24,2009. National conference on ICEEE09, Real time
simulation & testing by IEC61850
[6]Jan. 20-22,[Link] conference, ICAMTM-2006,
Next Generation-Plasma Based Power System
[7]Jan. 20-22,[Link] conference, ICAMTM-2006
SCADA-The Tool Of Development Of Economy
[8]Dec. 20-23, 2006. IEEE-International Conference, ADCOM2006 , Remote Control Of Process Parameters Using Embedded
Web Technology

Bibliography:-

Yogesh D. Sonawane is working with M/S Easun Reyrolle Ltd,

Mumbai as Asst. Manager with vast experience in power sectorEnergy Management & Automation. Published various papers at
National & International conference on Energy Management &
Automation.
Pankaj M. Patil is working with [Link] at Nashik as
Engineer-Testing with vast experience in power sector-Testing &
commissioning. Published various papers at national &
International conference on Energy Management & Automation.

V. CONCLUSIONS
An IEC 61850 messaging capability was
successfully developed for the RTDS Simulator. The
hardware implementation, referred to as the GTNET,
is capable of providing both binary input and output as
well as sampled value output for voltage and current
signals. Great care was given to ensure that the system
was deterministic in nature and able to operate error
free in conjunction with real time simulations. Both
the hardware and software used were optimized to
ensure that unacceptable delays were not introduced
when starting and stopping simulation cases. A
number of advantages were realized through the
development of the IEC 61850 messaging capability
for the RTDS Simulator. First it created the capability
for closedloop testing of newly compliant protection
devices. Second it greatly reduced the number of
physical connections that had to be made between the
simulator and the protection system. For the example
illustrated in Fig. 1 and 2, the number of connections
was reduced from 30 to just 4. Third, it was no longer
necessary to provide breaker status signals using
potential free contacts. Finally when using IEC 618509-2 sampled values, the need for power amplifiers was
eliminated, simplifying the test setup and greatly
reducing the cost.
VI. REFERENCES

Naresh D. Sonawane is presently working with M/S SIIRO Ltd,

Mumbai as Executive-Database Mgmt. with vast experience in
development and design software published various papers at
national & International conference on Automation & Software
support for application.
Sayali Y. Sonawane. Is student and doing Specialization in
Software designing and working on various software & Hardware
support to applications.

Websites:
[1] IEC 61850, located [Link]
Guides:
[2] UCA International Users Group, "Implementation Guideline
for Digital Interface to Instrument Transformers Using IEC
61850, R2-1.8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
174

Economical Benefits by Contribution of Large Wind

Farms to Voltage Control
A.-R. Al-Awaad, A.F. Kaptue Kamga, J.F. Verstege
Institute for Power Systems, University of Wuppertal,
Rainer-Gruenter-Strasse 21
42119 Wuppertal, Germany
The TSO forwards the cost based on EEG (EEG payment),
cost of delivery of reactive energy to support the voltage
control and the cost of cover of network losses to the
customers (Fig. 1). These costs are a part of network access
price.
Reactive energy price
Energy price to cover the network losses
TSO
Network access
price

Abstract- Wind energy is the main installed type of renewable

energies in Germany. Wind farms owner gets a fix payment for
the generated energy. This payment is much higher than the
average variable costs of conventional power plants. Wind farms
do not participate at voltage control till now. Although, they can
supply reactive power into the network. In this study it is
assumed, that there is a reactive power market. Large wind farms
are connected to high voltage networks. It is shown, that fed
reactive power from wind farms can contribute to minimization
of the network losses in high and very high voltage networks. This
fed reactive power leads to reduction of the complete demanded
reactive power to support the voltage control by strong load too.
Thereby, a part of costs of transmission system operator will be
saved. These saved costs can be a contribution to reduce the fix
payment of the generated energy of wind farms.

CPP
EEG payment

WF

customers

I.

INTRODUCTION

Today, renewable energy plays an important role in energy

supply. In some fields the development is still growing,
especially at wind energy. It is to be expected, that the installed
wind power in Germany until 2020 reaches 60 % of the
German total load [1].
In Germany the renewable energy is extra supported by a
governmental act (Renewable Energy Sources Act; EEG) [2].
Thus, wind farms owner gets a priority feed-in into the
network, where transmission system operator (TSO) must takeover the energy. Based on EEG and before the appearance of
wind energy trader, the TSO has to pay a fix payment (EEG
payment) to wind farms owner for the generated energy (Fig.
1). This payment is much higher than the average variable
costs of conventional power plants.
TSO is responsible for a sure operation of the electrical
network. Therefore, he has to support the voltage control [3].
TSO gets a part of the demand reactive power to support the
voltage control from the conventional power plants (CPP). In
this study it is assumed, that there is a reactive power market.
Thus, the TSO has to pay to the CPP owners the price of
delivery of reactive energy (Fig. 1).
TSO is responsible for energy transmission between power
stations and load. The transmission of the electrical energy
causes network losses. TSO gets the demanded active energy
to cover the network losses from CPP. TSO has to pay to CPP
owners the price of the demand active energy to cover the
network losses (Fig. 1).

Figure 1. Observed money flow

II. PARTICIPATION OF WIND FARMS AT VOLTAGE CONTROL

Modern wind power plants are implemented either with
double fed induction generator or synchronous generator in
combination with full-converter. These two types of wind
power plants can supply reactive power into the network [4, 5].
New planned large WF are to be connected to the high voltage
network [1]. These large WF do not participate at voltage
control till now. In this study it is investigated, which
economical benefits can be caused to the customers, by
reducing the access price, through participation of large WF at
voltage control.
III. TEST MODEL
The test model consists of a 380-kV- and two subordinate
110-kV-networks (Fig. 2 - 4). To the first 110-kV-network,
200 MW wind power are connected in four WF. These WF are
implemented with double fed induction generators. To the
second 110-kV-network, 700 MW wind power are connected
in eight WF. These wind farms are implemented with
synchronous generators in combination with full-converter.
Total strong load of the network is 4.2 GW. Utilization period
of WF is 1810 h/year and 6600 h/year of the load.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
175

G10
10

75 km
3

75 km

75 km

50 km

50 km
7

50 km
8

50 km
G9
27

50 km

75 km

14

34

75 km

75 km
50 km

11

75 km

12
75 km

42

Figure 2. 380-kV-network

13
30 km
20
30 km
WF

25

G7

25 km
14
20 km
17

20 km 15 km 25 km
16
WF
10 km
19
25 km
15 km 10 km
23
22
WF

25 km
15
WF
18

25 km
21
30 km
30 km
26

30 km
27

20 km

24
25 km

20 km

32
20 km
WF

10 km
15 km
36
10 km
WF 10 km
38
37
40
15 km
12
20 km
20 km
41
10 km
42
47
15 km
46
45
WF
48
10 km WF
31

43

WF

600
400
200

WF

35
15 km
10 km WF
15 km
39

15 km

800

30 km

20 km

34

1000

28

Figure 3. The first 110-kV-network

33

A. Strong Load & Weak Wind

By strong load, inductive reactive power is required to
support the voltage control. In this study, this necessary
reactive power is delivered from CPP at 380-kV-network. WF
are connected to 110-kV-networks. Thus, a part of the
necessary inductive reactive power for the voltage control can
be fed from these WF locally. Thereby, overhead lines and
transformers can be released from a part of the transmitting
inductive reactive power. Thus, the complete fed reactive
power of WF and CPP is less than the fed reactive power of
CPP alone (Fig. 5). E.g. the fed reactive power can be reduced
8 % approximately at the time between 11:00 and 12:00
oclock. As a result, with reactive power supply of WF the
network losses can be reduced too (Fig. 6). E.g. the network
losses can be reduced 6 % at the time between 11:00 and 12:00
oclock.
Reactive power supply (MVar)

50 km

100 km

G2

75 km

44

10 km
10 km WF
49
20 km

Figure 4. The second 110-kV-network

IV. TEST RESULTS

In this study, the two types of wind power plants (all
attached WF) supply reactive power into the network. The
network losses and the complete demanded reactive power to
support the voltage control are compared with and without
reactive power supply of WF. The network is tested for
different load and wind velocity cases. In this paper, only two
cases are investigated briefly: a week day in winter by weak
wind (strong load & weak wind) and a sunday in summer by
strong wind (weak load & strong wind). In this work, reactive
power supply are adjusted to get minimum network losses.

24

12
Time (h)
Without reactive power supply of WF
With reactive power supply of WF
Figure 5. Complete delivered reactive power

24

12
Time (h)
Without reactive power supply of WF
With reactive power supply of WF
Figure 6. Complete network losses

40
Network losses (MW)

G1
1

30
20
10
0

B. Weak Load & Strong Wind

Early in the morning by weak load and strong wind, the fed
active power of WF and CPP into the network is larger than the
load. WF have feed in priority. Thus, many CPP have to be
disconnected from the network to avoid a power surplus, e.g.
from 05:00 to 07:00 oclock the load is very low. Hence, CPP
G10, G2 and G7 have to be disconnected from the network.
The CPP, which remain at the network, have to produce the
necessary (capacitive) reactive power to support the voltage

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
176

control. Thus, a high reactive power flow can be caused and

the possibilities to minimize the network losses are limited.
Reactive power supply of WF offers more possibilities to
support the voltage control and to minimize the network losses
simultaneously. Thus, the complete delivered reactive power of
WF and CPP is larger than the complete delivered reactive
power of CPP alone in weak load (in the morning and evening,
Fig. 7). In this case, WF may supply reactive power, only if the
saved costs of network losses are larger than the more costs of
delivery of reactive power. Through reactive power supply of
WF, the network losses can be reduced, e.g. 5 % approximately
at the time from 11:00 to 12:00 oclock (Fig. 8).

Reactive power supply (MVar)

TABLE I
PRICES OF REACTIVE ENERGY DELIVERY TO SUPPORT THE VOLTAGE CONTROL
Price (/MVarh)
Reactive
Energy

Inductive

Capacitive

0.5

TABLE II
PRICE OF ACTIVE ENERGY DELIVERY TO COVER THE NETWORK LOSSES

600

Price (/MWh)

500

HT

400

Voltage
Level

300
200
100
0

12
Time (h)

24

Without reactive power supply of WF

With reactive power supply of WF
Figure 7. Complete delivered reactive power

16
Network losses (MW)

losses (0.52 m) and a part of costs of delivery of reactive

energy (0.13 m). WF owner gets more income (1 m) through
delivery of reactive energy, because there is a reactive power
market.

380 kV
110 kV

LT

50

50

80

60

D. Economical Benefits for Customers

In this study, the EEG payment is 87 /MWh. In addition to
that, it is assumed, that the total income of WF owner and costs
of TSO with and without participation of WF at voltage control
may not be changed. Therefore, the more income of WF owner
and the saved costs of TSO should be transferred to the WF
owner as a part of the fix payment (EEG-payment) for the
observed year. This leads to reduction of the EEG payment.
Thus, the EEG payment can be reduced from 87.00 to
85,98 /MWh (-1,17 %) in the observed year (Fig. 9).

12
8
0.37 %

4
0

0.7 %

0
+

12
Time (h)
Without reactive power supply of WF
With reactive power supply of WF

0.1 %

24

Figure 8. Complete network losses

C. Economical Benefits by Contribution of Wind Farms to Voltage

Control for one Year
Tables I and II show the prices of reactive energy delivery to
support the voltage control and the prices of active energy
delivery to cover the network losses. High tariff (HT) prices by
active energy delivery have to be paid from 08:00 to 20:00
oclock on a week day, else the delivered energy should be
paid in low tariff (LT).
Through the contribution of WF at voltage control for one
year, the TSO saves a part of costs of covering of network

Additional income of WF owner though reactive power supply

Saved reactive power costs of TSO
Saved costs of TSO by network losses
+

Possible reduction of EEG payment

Fig. 9. Possible reduction of EEG payment in the observed year

EEG payment should be paid for 20 years. This payment is

not constant for the 20 years, it has a degression. Thus, this
payment was 83.76 /MWh in the year 2007 [2], and the
delivered energy of wind farms in Germany reached 39500
GWh in the same year [6]. So based on EEG, customers should

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
177

pay for this generated energy 3.3 b. The reduction of EEG

payment with 1.17 % means that, customers should pay 56 m
less.

of wind power plants to frequency control leads to reduce the

EEG payment too. This reduction is clearly larger than the
reduction by participation of wind power plants to the voltage
control.

V. CONCLUSION
In this study it is assumed, that there is a reactive power
market. Wind farms owner gets additional income through
participation of wind farms at voltage control. This
participation leads to reduction of network losses and the
demanded reactive power to support the voltage control by
strong load. This means, the transmission system operator
saves a part of the demanded costs to cover the network losses
and delivery of reactive power to support the voltage control.
The additional income of wind farms owner and the saved
costs of transmission system operator can be a contribution to
reduce the EEG payment, where wind farms owner and
transmission system operator should have the same income and
costs with and without participation of wind farms at voltage
control. Thus, customers pay less for energy of wind farms.
The new research should help to reduce the EEG payment of
wind power plants. The new researches show, that wind power
plants can support the frequency control too [7]. Contribution

REFERENCES
[1]
[2]

[3]
[4]
[5]

[6]
[7]

Planning of the Grid Integration of Wind Energy in Germany Onshore

and Offshore up to the Year 2020 (dena Grid Study), German Energy
Agency, Berlin, 2005
Act on Granting Priority to Renewable Energy Sources (Renewable
Energy Sources Act), Federal Ministry for Environment, Nature
Conservation and Nuclear Safety, Federal Law Gazette I, No. 40, Bonn,
2004
TransmissionCode 2007 Network and System Rules of the German
Transmission System Operators, Berlin, 2007
A.-R. Al-Awaad, J. Verstege, PQ-Curve of Wind Power Plants,
European EMTP-ATP Conference, Len, Spain, 2007, pp A1
A.-R. Al-Awaad, J. Verstege, Contribution of Wind Energy Plants to
Voltage Control in High and very High Voltage Networks, First
Electrical Engineering Conference, University of Aleppo, Aleppo, Syria,
2007, pp PS04
Development of Renewable Energy Sources in Germany in 2007,
Federal Ministry for Environment, Nature Conservation and Nuclear
Safety, Berlin, 2007
A.-R. Al-Awaad, S. Vller, J. Verstege, Contribution of Wind Power
Plants to Frequency Control, 4th Energy and Exergy Conference,
American University of Sharjah, Sharjah, United Arab Emirates, 2009, in
press

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
178

Optical properties of thin films of polyazomethine

with triphenylamine unit in the main chain prepared
by spin-coating method
M. Palewicza,b, A. Iwanb, J. Doskoczc, W. Strkc, D. Skd
a

Institute of Electrical Engineering Fundamentals, Wroclaw University of Technology, Wybrzeze Wyspianskiego 27

50-370 Wroclaw, Poland
b
Electrotechnical Institute Division of Electrotechnology and Materials Science, M. Sklodowskiej-Curie 55/61, 50-369 Wroclaw,
Poland
c
Institute of Low Temperature and Structure Research, Polish Academy of Sciences, Okolna 2, 50-422 Wroclaw, Poland, P.O.
Box 1410, 50-950 Wroclaw 2, Poland
d
Centre of Polymer and Carbon Materials, Polish Academy of Sciences, 34 M. Sklodowska Curie Street, 41-819 Zabrze,
Poland

Abstract-Thin films of a polyazomethine (PAZ) were prepared

from a chloroform solution by spin-coating technique. The
solution of the polymer was spread on microscopic cover glass and
quartz under nitrogen atmosphere at room temperature and
atmospheric pressure. The optical properties of the PAZ were
investigated at room temperature under atmospheric pressure. An
absorbance, transmittance and reflectance were determined from
spectrophotometric measurements. The absorption coefficients
and bang gap of films material were developed. Surface structure
of the PAZ thin layer was checked by AFM measurements.
Thickness of the PAZ thin layer on the glass and quartz were
determined with ellipsometer at the range from 150 to 220 nm.

I.

azomethine moieties (HC=N) in TPA may lead to new

functional materials based on the synergistic effect of both of
them. Additionally, polyazomethines are known to display
interesting thermal, mechanical, electrical, optical and fiberforming properties and they are promising materials in
optoelectronic and photonic application [3].
PAZ with TPA in the main chain can be used as one of the
layers in organic solar cells. The aim of this article is to
describe optical properties and texture of thin layer composed
from PAZ by spincoating method.

INTRODUCTION

Investigation of a new kinds of organic materials for

photovoltaic devices were begun because of increasing demand
of logging energy. Solar panels produced hitherto on industrial
scale are based on inorganic crystals like silicon. However,
manufacturing of such devices are still expensive and has
negative effect on natural environment. For these reasons an
organic materials for photovoltaic devices are investigated. Use
of the organic materials are supposed to be very promising in
photovoltaic devices because of economic reasons and high
demand of a products [1]. One of the first solar cells which
could be considered as organic origin was characterized by
Tang [2]. A thin-film, two-layer organic photovoltaic cell was
fabricated from copper phthalocyanine and a perylene
tetracarboxylic derivative [2]. However, no product of
conducted research has been used in commercial application.
Among them, much attention is paid to triphenylamine
(TPA) derivatives because they are promising candidates for
photo- and electro- luminescence materials. TPA is a unique
molecule possessing useful function such as redox-activity,
fluorescence, and transport of positive charge centers via the
radical cation species. Being efficient hole conductors TPA are
commonly used us photoconductors. Introduction of

II. EXPERIMENTAL
PAZ was prepared from dialdehyde and diamine via high
temperature solution polycondensation. Properties of the PAZ
were confirmed by elemental analysis, FTIR and 1H, 13C NMR
spectroscopy and are described in [4]. Structure of the PAZ is
shown in Fig. 1. The absorption spectrum of the PAZ in
chloroform solution was characterized by one well defined
band at 413 nm being responsible for -* transmission in the
imine group. PAZ emitted blue light in chloroform solution
with maximum of emission band at 480 nm under 400 nm
excitation wavelength [4].
Solution to spread on substratum was prepared from PAZ
and chloroform. Quantity of the used polyazomethine in solid
state and chloroform was 36 mg and 2 ml, respectively.
Dissolution of the PAZ was executed at room temperature.
H
O

H3C

O
H
N
N

H
O

n
CH 3

Fig. 1. Chemical structure of the PAZ.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
179

Basis from microscopic cover glass and quartz were situated

in ultrasonic washer with deionised water for ten minutes
period and then were cleaned in solution of deionised water
and 2-propanol (1:1 relation), toluene, lavaged in deionised
water and acetone [5].
The thin layers of the PAZ were spread by spin-coating
method via drift process on the glass and quartz. To avoid a
contamination and defects in thin stratum this process was
carried out at room temperature in atmosphere of nitrogen
under atmospheric pressure. The values such as rotational
speed and time of turn were set on 880 turns per minute by 10
seconds. Two beam spectrophotometer Cary 5000 was used
to measure optical properties of the thin PAZ layer. The
absorption (A), transmission (T) and reflectance (R)
measurements were performed at room temperature.
Topography of the surface was tested by an atomic force
microscope (AFM) in the contact mode. Thickness (d) of the
layer was measured using ellipsometer at constant angle.

A( E

EG )r

(2)

Where is absorption coefficient, A is parameter

independent of photon energy, E, EG represent photon and band
gap energy, respectively. Index r accepts two values 1/2 or 2
connected with direct and indirect band to band transition.
Type of the transition in the investigated material is determined
by subscript r = 2. Such approach leads to equation (3).

( E )1 2

f(E)

(3)

Delimitation value of the energy gap (EG) is determined from

linear approximation of equation (3), as is mentioned in [8] for

III. RESULTS AND DISCUSSION

The absorption spectra of the PAZ film on the glass
(PAZ/Glass) and on the quartz (PAZ/Quartz) substratum are
shown in Fig. 2. Maximum of the absorbance was observed for
both samples at 405 nm. Also, on a spectrum related with layer
deposited on the glass was noticeable oscillation behavior
below 260 nm (See Fig. 2). That phenomenon is related with
crystalline structure of basis. Electromagnetic wave in
ultraviolet (UV) range was absorbed and converted into
phonon oscillation in background of the sample based on the
glass, while for light at that length quartz was transparent. To
describe dependence of the obtained thin layers form
substratum two kinds of them have been used. Besides of the
chemical purification quartz was polished. Therefore
appropriate prepared basis capacitate to obtained good quality
of the thin layer. From the analysis of the absorbance spectrum
it was possible to notice the variation of the thickness layer for
the glass and quartz, which was confirmed by elipsometric
measurements
From transmission (T) and reflectivity (R) measurements an
absorption coefficient is calculated according to the equation
(1) given in reference [6,7], where d is film thickness. The
typical absorption coefficient plot of the PAZ thin layer on
glass and quartz at entire energy scale from 1.2 to 6.0 eV is
illustrated in Fig. 3. An increase of the absorption coefficient
was observed on plot from energy value equal to 2.5 eV.

1
( 1 R )2
ln
d
T

Fig. 2. Absorption spectra of the PAZ film coated on the glass and quartz.

(1)

The absorption edge of the PAZ on the glass and on the

quartz was observed between 2.5 and 3.0 eV. For
semiconductor materials it is possible to calculate an energy
gap (EG) from relation given by formula (2) [8].

Fig. 3. Absorption coefficient of the PAZ thin layer on the glass and quartz.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
180

amorphous semiconductors. The absorption edge of the PAZ

thin layer on the glass and quartz is shown in Fig. 4. Value of
the energy gap for samples deposition on the glass and quartz
is elevated (2.45 0.02 eV).
Energy gap (EG = 2.45 eV) of the polyazomethine film on
the glass and quartz are similar. Insignificant changes into
value of EG related with amount of the solvent in solution were
observed. Increasing or decreasing of the band gap about 0.01
eV was caused by different concentration of the chloroform.
Characteristic broad and strongest section at the absorption
coefficient allocated at about 3.0 eV (Fig. 3) is due to interband transition between delocalized states which are related
with interaction of -orbitals in the polyazomethine [9].
Measurements of the PAZ thickness layer on the glass or
quartz substratum were performed with ellipsometer equipment
working at constant angle. Real refraction coefficients (n) of
the glass, quartz and polyazomethine layer were required to
conduct a measurements. Real refraction coefficient n = 1.42 of
the PAZ was estimated from equation (4) [10], where k, R are
imaginary refractive and reflection coefficient, respectively.

, (k

0)

TABLE I
THICKNESS OF THE POLYAZOMETHINE (PAZ) THIN LAYER
Code

Thickness of layer, d

[nm]

PAZ/Glass

200

218

203

208

211

PAZ/Quartz

155

162

155

157

154

films was observed by AFM. Characteristic granulation is

typical for the polymers thin layers and also agglomerates are
visible in Fig. 5. Scale of the granulated polymers
concentration are related with a porosity of the substratum and
technological parameters implemented to spin-coater
equipment.

(4)

Thickness of the thin PAZ layer on the glass and quartz was
changing from 150 to 220 nm (Table I). About 50 nm variation
of width between the PAZ spread on the glass and quartz were
noticed. Difference in thickness of the PAZ films are related
with used basis and technological parameters of the spincoating method.
Atomic force microscope (AFM) was used to describe
surface structure of the PAZ. Images of the PAZ on the glass
and quartz are represented in Fig. 5. Relatively homogenous
and smooth surface for the PAZ coated on the glass and quartz

Fig. 4. Absorption edge of the PAZ thin layer on the glass and quartz.

Fig. 5. AFM images (5 x 5 m2) of the a) PAZ/Glass and b) PAZ/Quartz.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
181

IV. CONCLUSION
Spin-coating method appeared to be an effective tool to
prepare good quality of the polyazomethine (PAZ) thin layer.
Optical properties such like the energy gap was obtained in
typical way for the amorphous semiconductors. The
technological parameters have influence on the energy gap,
thickness and surface structure of the investigated
polyazomethine. Presented results seem to be helpful to better
understanding of the physical properties of that
polyazomethine. It can be concluded that the investigated
polyazomethine is an amorphous semiconductor with the wide
energy gap. Therefore such polyazomethine with the TPA in
the main chain could be used to produce photovoltaic devices.
REFERENCES
[1]
[2]

Shajing Sun and Niyazi Serdar Sariciftci, Organic Photovoltaics:

Mechanisms, Materials and Devices, Published in 2005 by CRC Press.
C. W. Tang, Two-layer organic photovoltaic cell, App. Phys. Lett. 48
(1986) 183-185.

[3]

A. Iwan, D. Sek, Processable polyazomethines and polyketanils: From

aerospace to light emitting diodes and other advanced applications,
Progress Polym. Sci., 33 (2008) 289-345.
[4] D. Sek, A. Iwan, B. Jarzabek, B. Kaczmarczyk, J. Kasperczyk, Z.
Mazurak, M Domanski, K. Karon, M. Lapkowski, Hole transport
triphenylamine-azomethine conjugated system: Synthesis and optical,
photoluminescence, and electrochemical properties, Macromolecules, 41
(2008) 6653 6663.
[5] J. A. Mikroyannidis, I. K. Spiliopoulos, A. P. Kulkarni, S. A. Jenekhe,
Synthesis and optical properties of poly(p-phenylenevinylene)s bearing
tetraphenylthiophene or dibenzothiophene moieties along the main
chain, Synth. Met. 142 (2004) 113120.
[6] J. Misiewicz, Podstawy optyki ciaa staego, Oficyna Wydawnicza
Politechniki Wrocawskiej, Wrocaw 1996.
[7] G.I. Rusu, A. Airinei, M. Rusu, P. Prepelit, L. Marin, V. Cozan, I.I. Rusu,
On the electronic transport mechanism in thin films of some new
poly(azomethine sulfone)s, Acta Materialia 55 (2007) 433442.
[8] B. Jarzabek, J. Weszka, M. Domanski, J. Jurusik, J. Cisowski, Optical
properties of amorphous polyazomethine thin films, Journal of NonCrystalline Solids 352 (2006) 16601662.
[9] J. Sanetra, Efekt fotowoltaiczny w organicznych ogniwach sonecznych
wybrane zagadnienia Monografia, Politechnika Krakowska im.
Tadeusza Kociuszki, Krakw 2006.
[10] B. Jarzbek, J. Weszka, A. Burinan, G. Pocztowski, Optical properties
of amorphous thin films of the Zn-P system, Thin Solid Films 279
(1996) 204-208.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
182

A new approach to a solar cell problem

Bronislaw Swistacz
Wroclaw University of Technology
Wyb. Wyspianskiego 27
Wroclaw, Poland
Abstract- In this paper a new concept for a space charge
transport through a solid placed between the two electrodes is
presented. The effect of the light on the electric field distribution
and on the shape of current voltage characteristic is determined.
For a space charge distribution some new singular solutions are
obtained. In this paper it is found that the system can act as a
blocking diode.
Keywords: double injection, space charge, trapping levels.

I.

INTRODUCTION

One of the fundamental problems of a macroscopic theory of

electric conduction is to find the total concentration of charge
carriers in a solid placed between the two electrodes. In this
paper, we will assume that the divergence of the electric field
distribution
will be defined by the total concentration of
carriers. Also, we will suppose that the contact processes have
an influence on the shape of the electric field distribution,
which corresponds to a current voltage characteristic.

processes. When an external electric field supplies the different

kinetic energy to the two adjacent atoms the trapped electron
can pass from trap to trap in the given trapping level. When the
external electric field is applied the electron can pass from the
cathode into a solid. Moreover, this electron can become free.
Similarly, the hole injection occurs when the valence electron
can pass from the bulk into the anode and the valence state is
empty. When the external electric field gives a portion of
kinetic energy to the valence electron of an adjacent atom, this
electron can pass from the atom to the given atom and can fill
the empty state on the valence level. The metal-solid-metal
system will be represented by a planar capacitor system
( Fig.1). With these above assumptions we can find the electric
field distribution E (x) and a current-voltage characteristics
J (V ) , here, J is the current density and V is the applied
voltage. In order to find these functions, we have to define the
boundary functions describing the mechanisms of carrier
injection from the anode and the cathode into the bulk [1-7].

The purpose of this work is to find a shape of a current-voltage

characteristic J (V ) of the metal solid metal system.

the bulk
the metal

2. THE MODEL
In this paper, for an orbital electron in the given atom, we will
assume that the total energy (the sum of the positive kinetic
energy and the negative potential energy of the electric field of
the positive nucleus) is negative and that the zero reference
level is at a finite distance from the nucleus. In the case of an
isolated atom the total energy of an orbital electron is negative
for any distance from the nucleus. When an orbital electron
absorbs a portion of the kinetic energy of a photon or a phonon,
the total energy of the electron increases. The inverse case is
when an orbital electron can lose a portion of the total energy.
This is caused by the Coulomb force interaction between the
orbital electron and the positive nucleus. Let us take into
account the two valence electrons of an isolated atom. When
this atom is packed into a solid, these electrons can occupy the
higher energy state. Next, when a portion of the kinetic energy
is given to the two valence electrons, these electrons can
become free. The two empty energy states (which are left by
the electrons) represent the two holes. Between the zero and
valence levels many energy states (the trapping states) are
available for the holes and for the electrons. When an orbital
electron absorbs a photon, this electron can pass from the
valence level to the zero level via trapping levels. The inverse
case is when an orbital electron can pass from the higher
trapping level to the lower trapping level and an energy portion
is emitted Analogously, we are known as the allowed
transitions for the trapped holes. Such the allowed electron
hole transitions are called carrier generation recombination

the metal

R the external resistance

qv

qv

the space charge density

VS

the source voltage

_
VS
Fig.1 A planar capacitor is connected with an external
resistance and a voltage source. Here, (+ -)
denote the
terminals of a voltage source.
+

3. DISCUSSION and CONCLUSIONS

When carrier generation processes are dominant, the electric
field intensity satisfies a homogenous equation. Upon these
conditions there exist two singular particular solutions as well
as the general integral. Additionally, when the boundary
functions are strongly increasing, we ascertain that there exists
a set of values of applied voltage V in which the current
density is not defined. This property denotes that the whole
system acts as a solar cell, Also, we see that the J (V ) curve
is strongly increasing and there can be J (V ) 0 . Therefore,
the system acts as a perfect blocking diode and a voltage
stabilizer. Also, this mathematical property corresponds to a
solar cell.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
183

REFERENCES
[1] M. A. Lampert and P. Mark : Current injection in solids. Academic Press,
New York, 1970.
[2] K. C. Kao : Double injection in solids with non-ohmic contacts: II. Solids
with defects. J. Phys. D.: Appl. Phys., vol. 17, 1984, pp. 1449-1467.
[3] B. wistacz : Carrier generation and the switching phenomenon. Further
theoretical description. J. Phys.: Condens. Matter, vol. 7, 1995, pp. 10037-1008.

[5] B. wistacz : Some further description for a current flow through an

amorphous solid. A case of an imperfect contact. Electron. Tele. Quart., vol. 53,
No. 2, Warsaw, 2007, pp. 143-154.
[6] B. wistacz : Bipolar space charge problem for semiconductors and
insulators. J. Phis.: Condens. Matter, vol. 7, 1995, pp. 2563-2585.
[7] B. wistacz.: A bipolar space charge problem for solids including a
secondary electron emission. Electron. Tele. Quart., vol. 53, No 3, Warsaw,
2007, pp. 309-326.

[4] J. Simon and J.-J. Andre : Molecular Semiconductors. Photoelectrical

Properties and Solar Cells. Berlin, Springer, 1985.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
184

Proposed Techniques for Identifying Faulty Sections

in Closed-Loop Distribution Networks
Wael Al-Hasawi

Mahmoud Gilany
College of Technological Studies,
Electrical Technology Dept.,
Kuwait

Abstract This paper deals with three problems in closed

loop distribution systems. The first problem is how to detect
and identify the faulty section after a short-circuit fault is
successfully isolated by the protection system. The second
problem is how to identify a section affected with undetected
open-circuit fault. A novel circuit is suggested in this paper to
solve these two problems. The third problem addressed in this
paper is the undetected short-circuit fault which may lead to a
complete shut down of the substation. An effective
modification in the trip circuits of the relays protecting the
main feeders of the closed loop is suggested to minimize the
occurrence of this problem. The validity of the proposed
circuits is checked against the specifications of the available
equipment used in the network.

Artificial Intelligent (AI) such as Artificial Neural Network

(ANN). A brief review of some of the fault location
techniques can be found in [3]. Most of the ANN based fault
location techniques relied on the information about the status
of circuit breakers and relays. A brief comparison of various
analytical techniques with ANN in transmission system fault
location is provided in [4]. Artificial neural networks, when
applied directly to fault diagnosis problem utilizing the time
variation of fault current as input signal, suffer from the large
CPU time required for the training and also dimensionality of
the network. Hence, some preprocessing technique is required
to reduce input data set.
The Wavelet Transform (WT) theory provides an
effective way to examine the features of a signal at different
frequency bands. These features may be essential for pattern
Index Terms-- Distribution networks, open-circuit faults, recognition. Hence, it is well suited for the fault identification
Fault identification, fault detection.
and classification in the power systems [5-6].
For open circuit fault detection, there are different
techniques. Most of these techniques are designed for
I. INTRODUCTION
istribution systems being the largest portion of overhead distribution networks. One of the algorithms used to
the whole network, diagnosis of faults becomes a detect open conductor is the technique developed by Lee and
challenging task. Faults in distribution systems effect power Bishop from Pennsylvania Power and Light (PP&L) in USA
system reliability, security and quality. Accurate fault location [7]. They developed a prototype Ratio Ground Relay for the
minimizes the time needed to repair damage, restore power detection of broken conductors. This relay depends on the
and reduce costs. The application of traditional fault location ratio setting between the zero sequence current components
techniques that use fundamental voltages and currents at line and the positive sequence current component.
Another algorithm is developed to identify and locate
terminals for distribution lines with tapped loads is difficult
downed
conductor fault case by monitoring the voltage
[1].
unbalance
along the distribution feeder. As the system
A fault or a disturbance, which leads to high values of line
measures
the
unbalance voltage at various points so it could
currents, is generally detected by the protective devices and
give
indication
of the fault place [8].
faulty section is isolated using re-closures and/or circuit
Considering
the extensive size of the network, these tasks
breakers. However, the location of the fault and identification
can
be
effectively
achieved through implementing systems
of the fault are normally not known.
utilizing
the
available
high-speed computer and
The system restoration can be expedited very fast if the
communication
technology.
The
Institute of Electrical and
location of fault is known or can be estimated to some
Electronic
Engineers
(IEEE)
has
defined Distribution
accuracy. Hence, faults in a distribution system have to be
Automation
System
(DAS)
as
a
system
that
enables an electric
detected instantaneously, irrespective of whether they are of
utility
to
remotely
monitor,
coordinate
and
operate
distribution
permanent or temporary nature, to isolate only faulty section.
components,
in
a
real-time
mode
from
remote
locations
[9].
Identifying the fault section, forming part of fault diagnosis
The
DAS
is
based
on
an
integrated
technology,
which
aiming at minimizing the maintenance and repair time.
In the last few decades, considerable amount of work has involves collecting data and analyzing information to make
been done in the area of fault diagnosis particularly to the control decisions, implementing the appropriate control
radial distribution system. The techniques used with systems decisions in the field, and also verifying that the desired result
fed from two ends or ring systems are very limited since the is achieved. The software acquires the system data (both static
and dynamic) and converts it into an information system. The
protection of such systems is more complicated [2].
Many standard techniques are based on algorithmic engineering analysis software provides the control decision
utilizing
the system
information.
The 10-13,
decision
8. EEEIC
Conference
on Environment
Engineering,
Karpacz,
Poland, May
2009making feature
approaches but
some International
latest techniques
involve
the use and
of Electrical

185

of the distribution automation distinguishes it from the normal

Supervisory Control and Data Acquisition (SCADA) system
[10]. However, such systems may not be accepted because of
its high expenses.
This paper takes part of the Kuwait distribution system as
a case study. It deals with three types of problems related to
closed loop distribution systems.
Two problems are investigated in Section III. The first
problem is how to identify successfully the faulty section after
a short-circuit fault is successfully detected and isolated by the
differential protection system. Under such condition, the
affected section is not easily identified. The second problem is
related to permanent open-circuit fault which is not detected
by any protection system and hence its section is not also
identified. A novel circuit is suggested in section III in order
to solve these two problems.
The third problem addressed in the paper is related to the
permanent short-circuit fault which is not detected by neither
the differential relay nor the overcurrent relay, but isolated by
the earth fault relay of the high voltage-side. Such a fault
results in a complete shut down of the substation. An effective
modification in the relays trip circuit is suggested to minimize
the occurrence of such a problem and it is covered in section
IV of this paper.
The performance of the proposed circuit is compared with
the equivalent Distribution Automation System (DAS). The
proposed circuit proves a similar performance as DAS with
minimum cost.
The problems studied are described by the electrical field
engineers and are investigated using ASPEN-OneLiner
simulation program, V9.7 [11]. ASPEN OneLiner is a PCbased short circuit and relay coordination program widely
used by protection engineers.

II. THE NETWORK UNDER STUDY

A typical 132/11 kV from Kuwait distribution network is
shown in Fig. 1. It contains closed loops with different sizes.
Every substation (represented with solid circle in the figure)
contains three 1000 kVA transformers. The distances between
the substations are shown on the drawing. The most right-hand
side closed loop is taken as a case study. Differential
protection is the applied protective system for every section of
the loop. Only the main feeders (F1, F2, and F3) outgoing
from the station are protected by additional overcurrent relays.
III. IDENTIFYING A FAULTY SECTION
The objective of this part of the paper is to fast identifying
the faulty section in a similar way like DAS but with a
minimum cost system. The key of the proposed circuits is to
get use of the spare wires of the pilot cable - associated with
the differential relays - to transmit certain information related
to the affected section to a PLC (Programmable Logic
Controller) unit in the main 132/11 kV station. Usually, the
pilot wire consists of about 16 pair of wires. In many cases,
only one or two pairs are used where the others are spare.
The input of the PLC for the studied network consists of
eight analog inputs. Information from a limited numbers of the
loop's CBs (and associated relays) are required. Only the
auxiliary output-contacts of the circuit breakers (and its
associated relays) represented with solid rectangular in Fig. 2
are needed in the proposed circuit. The eight input signals are:
four "Normally-Close" (NC) output contacts from the CBs
represented with solid rectangular plus four auxiliary
"Normally-open" (NO) output contacts of its associated relays.

Fig. 1 A typical Distribution Network

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
186

The Normally-close auxiliary output contact of any circuit

breaker will become "close" only if the circuit breaker's status
becomes "open". On the other hand, the auxiliary (NO) output
of the relay which is controlled by the relay algorithm will
become "closed" only if the current through the protected
feeder drops to zero provided that the CB contacts are still
closed.
The last condition is necessary to differentiate between
the zero current resulting from a normal feeder opening and
the zero current resulting from undetected open circuit fault.
The PLC unit processes the received information and
sends SMS message through a modem to a GSM cell phone.
The data included in this message informs the maintenance
team about the faulty section. The exact fault location within
the identified affected section is determined later - off lineusing any of the common used cable tracers.

Fig. 2: Identifying an open circuit section from both sides

A. Locating a Detected Short-Circuit Fault
In this case, we assume that there is a short circuit fault
which is detected and isolated by the differential protection.
The faulty section in this case will be opened from both sides
after clearing the fault. For example, for the fault occurred in
the section between substations H75 and H53 as shown in Fig.
2, the differential protection system detects this fault and
isolates the affected feeder. The two circuit breakers: CB5 and
CB6 will then be opened after clearing the fault.
The costumers will not be affected by opening the faulty
feeder since the system is originally fed from more than one
point. The problem is that the utility maintenance teams will
not also feel that there is a fault that has occurred and cleared
in that loop or in that section. The only available information
which may be useful to attract the attention of the maintenance
team to that loop is the disturbance in current distribution in

the three main feeders (F1, F2 and F3) shown in Fig. 2. These
currents are monitored by the supreme control center. The
maintenance team usually depends on this sole information
current disturbance - to predict the location of affected loop
and the opened feeder. In many cases, this variation in the
current distribution doesn't give a clear indication about
neither the affected loop nor the affected section. The
maintenance teams in many cases have to search for it from a
substation to another.
In the proposed circuit, the status of the "normally-closed
auxiliary contact" of the selected CBs is transmitted to the
PLC unit at the substation through the spare wires of the pilot
cables. For example, the status of the auxiliary contact of CB5
(see Fig. 2) is transmitted to the main station M using spare
wires of two pilot cables: a spare pair from the cable between
"H75 and H33" and then the spare cable between "H33 and the
main station, M". Only a junction between the two spare wires
is required to be added to facilitate this operation.
Once a fault is cleared, the status of the NC auxiliary
contact of CB5 will be changed from "open" to "close".
Consequently, SMS message will be forwarded to a certain
cell phone number stored in the PLC program. The
maintenance team is easily informed about the cleared fault
and hence the faulty section is identified. Any other detected
short circuit fault is identified in a similar way.
An alternative method for identifying the faulty section
needs an RTU in each point in the loop plus an efficient
communication network between all the load points. Other
alternative techniques which depend on calculating the fault
distance are not easily implemented with closed loop systems.
The main feature of this technique is to instantaneously
identify the faulty section like DAS - but with minimum
additional cost added to the existing system (the cost of only
one PLC unit per substation). It saves a lot of time and efforts.
B. Locating an Undetected Open-Circuit Fault
Open conductor (downed conductors) from the point of
view of distribution utility is a public hazard in the main
consideration. It is not a system operation problem since the
system could continue without disconnecting such fault
[12,13].
The second fault scenario studied in this paper is to have
an undetected open circuit fault. This may happen as a result
of a cut in the power cable. In this case, the power cable (and
may be the pilot cable as well) is opened but the circuit
breakers of the faulty section - at both sides - are kept closed.
The only available indication showing that there is a
problem is the disturbance in currents distribution in the three
main branches of that loop (F1, F2 and F3 in Fig. 2).
However, this information is not guaranteed as the variation in
the currents may be very small. It is not always easy to
identify the affected section based on these current readings.
Again, even if we got information about the disturbance in
certain loop, this information can't tell us about the faulty
section in the affected loop specifically. The maintenance
crew has to go through the stations to check the status of the
breakers. Inspecting the location of faults is done with manual
intervention and are rectified in a time consuming way. For

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
187

some faults, it takes a long time and a lot of costs to solve the
problem.
With the proposed circuit under such condition, the PLC
unit will receive a signal from the NO auxiliary contact of the
relay associated with the faulty section once the current in that
section drops to zero and provided that the status of the CBs is
not changed. The case is then identified as "undetected opencircuit fault".
For example, if undetected open circuit fault is assumed
on the section between stations H38 and H53, then the NC
auxiliary contact of CB-8 will be kept "open" while the
auxiliary NO contact of relay-8 will be changed to "close"
status.
IV. AVOIDING UNNECESSARY COMPLETE SHUTDOWN
The third problem addressed in the paper is to have a
permanent short circuit fault which is not detected by any of
the protection systems in the low-voltage side (neither
differential nor overcurrent protection) but it is isolated by the
transformer standby-earth fault relay located at the high
voltage side. In this case, not only the faulty loop will totally
be disconnected but also all the other healthy loops supplied
from the same bus bar at the substation (complete shut down).
There several reasons for such a problem. In some cases,
there is a poor discrimination between the 132/11 kV
transformer primary overcurrent relay and overcurrent relays
protecting the outgoing feeders. A typical case is shown in
Fig. 3.
The differential protection can't detect such a fault as it
occurs directly on the bus-bars of substation H53. The phase
overcurrent protection (OP's) for the three main feeders will
all trip to isolate the fault.
It can be seen from Fig. 3 that there is a very short time gap
between the transformer primary OC relay operation (2.71
sec) and the relay of feeder FD1 (2.69 sec.). There is a degree
of uncertainty in knowing which one will trip first. In many
cases, both operate at the same time. This coordination
problem is one of the reasons that lead to unnecessary
complete shutdown of the station.
In some other cases, especially with a single line to
ground fault, there is a probability that one of the three
overcurrent relays (F1, F2 and F3) may fail to detect the fault.
F3

OP 2.71s

This case is also expected if the fault occurred while only two
of the three 132/11 kV transformers are in service or if the
fault occurred through a fault resistance. Under such
conditions, the sensitivity of the relay will be reduced. The
fault current will find a path to the fault point even after
disconnecting the other two relays. If this fault persists for
long time, the transformer standby earth fault protection installed in the transformer neutral connection - will trip the
main transformer.
Usually, all the three 132/11 kV transformers are
connected in parallel with common standby earth fault relay. It
means that the station will be completely shut down under
such conditions.
The principle of the proposed solution is based on the fact
that if a fault in any section is detected by the corresponding
differential relay, then none of the overcurrent relays R1, R2
and R3 (see Fig 4) will operate. On the other hand, closing the
contacts of any of the three overcurrent relays R1 or R2 or R3
means that the differential relay for a certain section failed to
operate.
The idea of the proposed solution is to accelerate the trip
of the other two loop breakers as soon as a fault is detected by
any one of the three OC relays. The three CBs will trip
simultaneously once any relay contacts is closed as shown in
Fig. 4. This will partially prevent the condition of feeding a
fault from un-tripped feeder and consequently avoid the
complete shutdown.
V. CONCLUSIONS
This paper presents a practical field experience with
Kuwait distribution networks. Novel circuits for identifying
the faulty section in case of detected short circuit fault and
undetected open-circuit fault are presented. Another circuit is
presented to avoid the problem of shut down of the whole
132/11 KV station. The proposed-circuits reduced technical
and commercial losses, lower electric service restoration time,
reduce the equipment damage, and enhanced power quality
and reliability. These circuits succeeded to fulfill many tasks
of DAS systems with almost neglected cost since it saves the
cost of communication network and the cost of the RTUs
required at each load point.

H33
OP 2.69s [Link]

F2 OP 1.83s

A8
[Link]
F1

OP 1.05s

H75
[Link]

A1
[Link]
A2
[Link]

H53
[Link]

H38
[Link]

13150P-87

H142
[Link]

A3
[Link]

A4
[Link]

HAWL E11 BS1

[Link]
A7
[Link]

B
[Link]

A5
[Link]

Fig. 3: Three line to ground fault at Bus-bar H53.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
188

[2]
[3]
[4]
[5]

[6]

[7]

Fig. 4: The modified trip circuit

ACKNOWLEDGMENTS
The financial support from the General Authority for
Applied Education and Training in Kuwait (Project No. TS02-08 ) is highly appreciated.
REFERENCES
[1]

Hassan Nouri, Chun Wang and Terry Davies, "An Accurate Fault
Location Technique for Distribution Lines with Tapped Loads", IEEE
Porto Power Tech Conference, Porto, Portugal, 10 -13 September, 2001,
pp. 488 - 493.

[8]
[9]
[10]
[11]
[12]
[13]

M. da Silvaa, M. Oleskoviczb, D.V. Coury, "A hybrid fault locator for

three-terminal lines based on wavelet transforms", Electric Power
Systems Research, Vol. 78, pp. 1980-1988, 2008.
M. M. Saba, R. Das, P. Verho, and D. Novosel, "Review of Fault
Location Techniques for Distrbution Systems," Power Systems and
Communication Infrastructure for the future, Beijing, Sep 2002.
G. K. Purushothama, A. U. Narendranath, D. Thukaram, and K.
Parthasarathy, "ANN Applications in Fault Diagnosis," Electric Power
System Research, vol. 23, no. 6, pp. 491- 506, Aug. 2001.
U. D. Dwivedi, S. N. Singh and S. C. Srivastava, "Analysis of Transient
Disturbances in Distribution Systems: A Hybrid Approach", 2007 IEEE
Power Engineering Society General Meeting , Tampa, Florida , USA,
24-28 June 2007.
M. Michalik, M. Lukowicz, W. Rebizant, S.-J. Lee, and S.-H. Kang,
High impedance fault detection in distribution networks with use of
wavelet-based algorithm, IEEE Trans. Power Del., vol. 21, no. 4, pp.
17931802, Oct. 2006.
R. E. Lee , M. T. Bishop , "Performance Testing of the Ratio Ground
Relay on a Four Wire Distribution Feeder", IEEE Trans. On PAS, Vol.
PAS-102, No.9, Sept ,1983.
Cesar Senger, "Broken Conductor Protection System Using Carrier
Communication", IEEE Transactions on Power Delivery, VOL. 15, NO.
2, April 2000.
R.P. Gupta and R.K. Varma, Distribution Automation: Present Status.
Academic Open Internet Journal, Volume15, 2005.
Musse M. Ahmed, "Electrical Distribution Automation System for Low
Voltage (LV) System", First International Power and Energy Conference
PECon 2006, November 28-29, 2006, Putrajaya, Malaysia.
ASPEN-OneLiner Simulation Program V9.7,[Link].
IEEE Guide for determining fault location on AC transmission and
distribution lines, IEEE Standard C37.114-2004, December 2004.
IEEE Tutorial Course Text, "Detection of Downed Conductors on
Utility Distribution Systems", 1989.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
189

Charcoals Used as Electrode Material in

Supercapacitors
B. Szubzda, R. Kulinski,
Electrotechnical Institute Division of Electrotechnology
and Materials Science
M. Sklodowskiej-Curie 55/61
Wroclaw, Poland
Abstract- In this paper we have been described biomimetical
structured charcoals based on acacia, sycamore, coconut and pine,
in which capacitance arises in consequence of separation of
charges in electrode/electrolyte interface (so-called electrical
double layer), due to it's biomimetical structure, are proven to
have high porosity and capacitance.
Influence of biomimetical structure on supercapacitive behaviour
has been tested by cyclic voltammetry, galvanostatic chargedischarge and self-discharge tests.

I.

INTRODUCTION

During whole history of mankind people were trying to draw

inspiration from the nature to make their lives better. In the
middle of XX century this tendency lead to rise of a new
interdisciplinary field of science biomimetics. So far a large
amount of new materials has been discovered e.g.: very light
biomorphic cellular ceramics with high thermal and
mechanical strength [1], artificial bones made from
hydroxyapatite [2], SiOC cellular ceramics used in biomedicine
and electronics [3], or lightweight, porous ceramics structures
for catalytic use [4, 5].
In recent years investigations on a new type energy storage
device, so-called supercapacitor, has been carried out
intensively. In supercapacitor high surface area of porous
carbon electrode and the phenomenon of electrical double layer
at electrode interface is used to obtain high capacitance, power
and energy density.
The natural structure of a living tree found to be very porous.
This natural porosity is connected with existence of special
tissues responsible for a transport of water and soluble mineral
nutrients from the roots to the rest of the plant. This natural
porosity can be used to obtain an interesting material for
supercapacitor electrode.
In this paper different types of pyrolyzed and activated
woods was tested as an electrode in supercapacitor.
II.

EXPERIMENTAL

A. Preparation of pyrolyzed and activated charcoals

Previously cutted out and prepared raw samples of four types
of trees: acacia, sycamore, coconut and pine were carbonized
in tubular furnace in controlled atmosphere of Ar gas for 2
hours at 1100 C. Then samples were activated by treating
solid KOH at 700 C for 90 min. The initial amount of

substrates in weight relation was 1.0 KOH / 0.25 wooden

substrate. Charcoals produced that way was then formed in
defined shapes and weighted.
B. Electrochemical tests
Previously cutted out and prepared, raw samples of charcoal
was tested as supercapacitor electrodes; 1M solution of H2SO4,
and 6M KOH were used as an electrolyte. All measurements
were made in five electrodes system using a device ATLAS
0531 Electrochemical Unit & Impedance Analiser. The
reference electrode was saturated calomel electrode.
The range of taken tests:
Cyclic voltammetry
Galvanostatic charging and discharging
Self-discharge tests
III.

RESULTS AND DISCUSSION

A. Cyclic voltammetry
Cyclic voltammetry measurements was carried out at a rate
of potential increase 1 mV/s to the value of 1.1 V, back to the
zero, and then for a negative potential equal -1.1 V and again
back to zero.
Analysis of the shape of voltammetric curve (CV) gives an
information about electrochemical phenomenon in system,
especially about electrodes phenomenon, resistance of system
and dynamics of electrical double layer charging and
discharging. The width of hysteresis loop corresponds to the
ability of electric charge accumulation and is a basis of
supercapacitor capacitance calculation. Fig. 1a, 1b, 1c, 1d
shows CV for different types of charcoal, before and after
activation in 1M H2SO4 and 6M KOH.
Activated charcoals are seemed to have wider hysteresis
loop, than charcoals before the activation, for both acidic and
basic electrolyte. Also the current response are considerably
higher than in charcoals before the activation.
B. Galvanostatic charging and discharging
Galvanostatic charging and discharging measurements are
very helpful in calculation of supercapacitor capacitance. To
charge and discharge the supercapacitor forced, constant

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
190

current was held and the time when potential between cell
electrodes reach previously fixed value was the variable. In this
case the current value was set as 1/5 of absolute value of
potential, e.g. if potential was set as 600 mV, current was 120
mA. The supercapacitor was charged and then discharged for
the following values of cell potential: 100, 200, 300, 400, 500
and 600 mV.
Results of galvanostatic charging and discharging
measurements was gathered in Table I. The capacitance of
supercapacitor was recalculated with respect to the mass of
electrode.
Activation of charcoals are supposed to have a great
influence on their capacitive bahaviour. In all cases the
capacitance is much higher than before the activation, e.g. pine
(Fig.1d), where the capacitance is more than hundred times
higher. Also it is worth to be noticed that in 6M KOH
electrolyte the capacitance is considerably higher than in acidic
one (with the exception of activated sycamore).
C. Self-discharge tests
There are many factors that have influence on the behaviour
of supercapacitor self-discharge, e.g. ionic mobility, sort and
strength of electrolyte, temperature, degree of solvation, factors
responsible for matching electrolyte and electrode: wettability,
size of electrode pores, and ions of electrolyte.
In measurements supercapacitor was charged to the potential
0,6 V at a rate of potential increase 10 mV/s. This potential was
forced for 10 minutes using external power source and then
was controlled for the next 11 hours.
The results of measurements was presented in Table II and in
Fig. 2.
Leakage current is very important parameter in selfdischarge analysis, it determines the amount of energy needed
to sustain the electrical double layer. Low leakage current
connected with high capacitance, as in activated acacia,
TABLE I
GALVANOSTATIC MEASUREMENTS - CAPACITANCE OF CHARCOALS IN ACIDIC AND
BASIC ELECTROLYTE

Sample

1M H2SO4

6M KOH

1 Acacia

27,9

2 Acacia - activated

79,9

170,6

3 Sycamore

58,8

4 Sycamore - activated

141,6

82,6

5 Coconut - activated

220,8

281,5

3,9

475,8

6 Pine
Figure 1. Diagram of cyclic voltammetry for (a) acacia , (b) sycamore, (c)
coconut, (d) pine; scan rate 1 mV/s.

Capacitance [F/g]

7 Pine - activated

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
191

TABLE II
SELF-DISCHARGE BEHAVIOUR FOR DIFFERENT KINDS OF CHARCOALS, IN ACIDIC AND
BASIC ELECTROLYTE

self-discharge after
charging to 600 mV
Sample

leakage
current
[mA]

potential
after 1 hour
[mV]

1 Acacia, H2SO4

4,1

155

2 Acacia activated, H2SO4

5,0

135

3 Acacia activated, KOH

1,6

379

4 Sycamore, H2SO4

4,0

437

5 Sycamore activated, H2SO4

9,3

246

6 Sycamore activated, KOH

3,0

422

7 Coconut activated, H2SO4

10,8

286

8 Coconut activated, KOH

3,1

529

9 Pine, H2SO4

0,2

104

10 Pine activated, H2SO4

10,1

360

Figure 2. Diagram of self-discharge of supercapacitor after charging to 600 mV

(1) acacia, H2SO4, (2) acacia activated, H2SO4, (3) acacia activated, KOH,
(4) sycamore, H2SO4, (5) sycamore activated, H2SO4, (6) sycamore
activated, KOH, (7) coconut activated, H2SO4, (8) coconut activated, KOH,
(9) pine, H2SO4, (10) pine activated, H2SO4.

coconut and sycamore) or hydrophobic properties (pine and

sycamore before the activation).
As a result of voltammetric measurements charcoals in
which in range of charging potentials unfavourable
electrochemical processes(decomposition of water) does not
exist was selected: coconut and pine.
REFERENCES

sycamore and coconut, is very desirable factor for

supercapacitors. It is worth to be noticed that in acidic
electrolytes leakage current is over three times larger than in
basic ones.
IV.

CONCLUSIONS

On the basis of the electrochemical measurements it was

concluded that the activation of charcoals is crucial for their
capacitance, charge, discharge and self-discharge behaviour.
For all samples supercapacitive properties are seemed to be
considerably greater than before the activation. The charcoals
obtained in this way appeared to have hydrophilic (activated

[1]
[2]
[3]
[4]
[5]

[6]

H. Sieber, Ch. Hoffmann, A. Kaindl, P. Greil, Biomorphic Cellular

Ceramics, Advanced Engineering Materials (2000), 2. No. 3, pp. 105
109.
K. Hata, T. Kokubo, Growth of a Bonelike Apatite Layer on a Substrate
by a Biomimetic Process, J. Am. Ceram. Soc.(1995), 78, 4, pp.
1049-1053.
P. Colombo, Novel Processing of Silicon Oxycarbide Ceramic Foams,
Advanced Engineering Materials, (1999) 1, No.3-4, pp. 203-205.
T. Ota, M. Imaeda, H. Takase, M. Kobayashi, N. Konoshita, T. Hirashita,
H. Miyazaki, Y. Hikichi, Porous Titnia Ceramic Prepared by Mimicking
Silicified Wood, [Link]. Cer. Soc. (2000) 83, 6, pp. 1521-1523.
T. Ota, Y. Hikichi, H. Unuma, M. Takahashi, H. Suzuki, H. Akahane,
Production of Ceramic Woods by Mimicking Fossil Woods, 9-th CimtecWorld Ceramic Congress, Ceramic Getting ito the 2000`s Part C
(1999), pp. 531-538.
B.E. Conway, Electrochemical Capacitors: ScientiJic Fundamentals and
Technological Applications, Kluwer AcademicPlenum, 1999.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
192

Wavelets for detection of voltage dips and micro

interruptions
Zbigniew Leonowicz

Wroclaw University of Technology

Wyb. Wyspianskiego 27
Wroclaw, Poland
Abstract- In electrical energy power network, disturbances can
cause problems in electronic devices therefore their monitoring is
fundamental in Power Quality field both to properly dimension
protections and to calculate compensations in case of malfunction
of the apparatus. In this paper we address the problem of
disturbances estimation by using Wavelets signal processing for
detection of short, impulse-like voltage dips and microinterruptions.
Index TermsPower Quality, Wavelets, voltage dips, micro
interruptions..

I.

INTRODUCTION

Electric power quality has became an important part of

power systems and electric machines, studied from a wide
number of points of view. For this purpose there are many of
electric parameters that help to describe the phenomena as a
whole are reported in standards. In this context, we consider
disturbances as the temporary deviation of the steady state
waveform caused by faults of brief duration or by sudden
changes in the power system [1]. The disturbances considered
by the International Electro-technical Commission include
voltage dips, brief interruptions, voltage increases, and
impulsive and oscillatory transients [2,3, 4].
The first ones are defined by norms as a sudden reduction
(between 10% and 90%) of the nominal voltage, at a given
point of electrical system, and lasting from half of the
fundamental period to several seconds. The dips with durations
of less than half a cycle are regarded as transients. The main
characteristics of voltage dips are magnitude and duration,
which correspond to the remaining bus voltage during the fault
and the required time to clear the fault respectively. A voltage
dip may be caused by switching operations associated with
temporary disconnection of supply, the flow of heavy current
associated with the start of large motor loads or the flow of
fault currents or short circuits and earth faults. These last ones
can be symmetrical (three phase) or non symmetrical (singlephase to ground, double-phase or double-phase-to-ground).
The brief interruptions can be considered as voltage sags with
100% of amplitude. The cause may be a blown fuse or breaker
opening and the effect can be an expensive shutdown. For
instance, supply interruptions lasting up to few seconds may
cost a lot in case of interruption of service or stoppage of
machines in a production plant. Costs that can quickly grow up
with the plant resetting time that can be very long. The main
protection of the customer against such events is the
installation of uninterruptible power supplies [1].

Brief voltage increases (swells) are brief increases in r.m.s.

voltage that sometimes accompany voltage sags. They appear
on the unfaulted phases of a three phases of a three-phase
circuit that has developed single-phase short circuit. Swells can
upset electric controls and electric motor drives, particularly
common adjustable-speed drives, which can trip because of
their built-in protective circuitry. Swells may also stress
delicate computer components and shorten their life. Possible
solutions to limit this problem are, as in the case of sags, the
use of uninterruptible power supplies and conditioners [5].
Voltage disturbances shorter than sags or swells are
classified as transients and are caused by sudden changes in the
power system [5]. According to their duration, transient
overvoltages can be divided into switching surge (duration in
the range of millisecond), and impulse spike (duration in the
range of microseconds). Surges are high-energy pulses arising
from power system switching disturbances, either directly or as
a result of resonating circuits associated with switching devices.
Protection against surges and impulses is normally achieved by
surge-diverters and arc-gaps at high voltages and avalanche
diodes at low voltages.
In this article we focus the attention on disturbances which
will gain more importance in the next future because of the
increase of electronic apparatus that can be particularly
sensible to this kind of problems if not adequately protected. In
fact there are two important aspects that should be taken into
account:
In all cases, in power quality is necessary to detect not only
the beginning and end of voltage sag but also to determine the
sag depth and the associated phase angle jump.
The performance of wavelet signal processing is tested for
analysis of short, impulse signals [8].
II. APPLIED WAVELET TRANSFORM ANALYSIS
Wavelet transform is a useful tool in signal analysis.
Wavelets provides a fast and effective way of analyzing nonstationary voltage and current waveforms and can be applied
for precise computation of the beginning of a disturbing event,
as shown in this paper. The ability of wavelets to focus on
short time intervals for high-frequency components and long
intervals for low-frequency components improves the analysis
of signals with localized impulses and oscillations, particularly
in the presence of a fundamental and low-order harmonic [6].
The continuous Wavelet Transform (WT) of a signal x (t ) is

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
193

X a ,b =

1
a

D2 level.

x(t ) (

t b
) dt
a

(1)

where (t ) is the mother wavelet, and other wavelets

1 t b
(2)

dt
a a
are its dilated and translated versions, where a and b are the
dilation parameter and translation parameter respectively,
a R + {0}, b R [7,8].
The discrete WT (DWT), instead of CWT, is used in
practice [23]. Calculations are made for chosen subset of scales
and positions. This scheme is conducted by using filters and
computing the so called approximations and details. The
approximations (A) are the high-scale, low frequency
components of the signal. The details (D) are the low-scale,
high-frequency components. The DWT coefficients are
computed using the equation
X a ,b = X j , k = x[n]g j , k [n]
(3)

300

v(t)

defined as

a ,b (t ) =

scale

0.8

symlet(32)

0.4

wavelet

Wavelet

2
1
0

Figure 2. Wavelet decomposition (lower plot) of the sinusoidal waveform

(upper plot) distorted by one voltage dip.

III. INVESTIGATIONS, DISCUSSION AND CONCLUSIONS

For experimental testing the performance of the algorithms,
we used a synthesized signal realized by MATLAB able to
generating voltage dips of different magnitudes.
For evaluating the performance of the method many test
signals has been used with different THD and SNR. The THD
used are 5.7%, 11.2% and 22.4%. These values were obtained
using for each of the first 24 harmonics half of the norm limits,
the norm limits, and the double of norm limits [9].
For each of the three THD has been created three signals
with a different SNR: 100dB, 80dB and 60dB. The added noise
is white Gaussian.

v(t)

200
0
-200
3

2
1
0

30

40

50

60

30

40

50

60

70

1
0
-1
20

Figure 3. Wavelet decomposition (lower plot) of the sinusoidal waveform with

harmonics (upper plot) distorted by one voltage dip.

0
20

-300

Wavelet

nZ

where a = 2 j , b = k 2 j , j N , k Z .
The wavelet filter g plays the role of [6]. The choice of
mother wavelet is different for each problem at hand and can
have a significant effect on the results obtained. Orthogonal
wavelets ensure that the signal can be reconstructed from its
transform coefficients .
As wavelet the symlets function was used. The symlets are
nearly symmetrical wavelets proposed by Daubechies as
modifications to the db family - orthogonal wavelets
characterized by a maximal number of vanishing moments for
some given support (Fig. 1). Dips detection was realized
through tracking values of details (D) representing higher
frequencies in the signal. High value indicated dip. In contrary
to other presented method this approach did not use the
amplitude parameter of the main component, but was therefore
prone to noise and other high frequency disturbances.

70

Figure. 1. Scale and wavelet symlets function for 32 coefficients

Figure 2 shows the behaviour of the wavelet decomposition

of the sinusoidal waveform distorted by one voltage dip. The
decomposition was made using the Daubechies 6 wavelet at the

Higher order frequency components present in the signal

deteriorated the detection ability of the method. Two wavelets
with significantly different lengths have been used; Symlet
(length of the filter 32 samples) and Daub 6 (length of the filter
6 samples).
In the nine test signals 100 dips have been added, one for
each period in randomly position. In Tables I and II are
reported the percentage of detected dips for 35V and 100V dips
of respectively 15% and 43% of nominal voltage.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
194

TABLE I
DETECTION ACCURACY OF 35V DIPS

SNR
THD

100 dB

80 dB

60 dB

5.7 %

100 %

45 %

12 %

11.2 %

100 %

45 %

17 %

22.4 %

100 %

33 %

14 %

TABLE II
DETECTION ACCURACY OF 100V DIPS

SNR
100 dB

80 dB

60 dB

5.7 %

100 %

100 %

70 %

11.2 %

100 %

100 %

67 %

22.4 %

100 %

100 %

62 %

THD

REFERENCES
[1] J. Arrillaga, N.R. Watson, S. Chen, Power System Quality Assessment,
John Wiley & Sons, 2000, ISBN 0 471 98865 0.
[2] IEC 61000-2-1: 1990, Electromagnetic Compatibility (EMC), Part 2:
Environment, Section 1: Description of the Environment for LowFrequency Conducted Disturbances and Signalling in Public Power
Supply Systems.
[3] IEC 61000-2-2: 1990, Electromagnetic Compatibility (EMC), Part 2:
Environment, Section 2: Compatibility Levels for Low-Frequency
Conducted Disturbances and Signalling in Public Power Supply Systems.
[4] IEEE c62.41: 1991, IEEE Recommended Practice on Surge Voltages in
Low-Voltage AC Power Circuits.
[5] UIE-DWG-2-92-D, UIE Guide to Measurements of Voltage Dips and Short
Interruptions Occurring in Industrial Installations.
[6] S. Santoso, E.J. Powers, W.M. Grady, P. Hofmann: Power Quality
Assesment via Wavelet Transform Analysis, IEEE Transactions on
Power Delivery, 11(2), pp. 924-930.
[7] M. Misiti, Y. Misity, G. Oppenheim, J-M. Poggi, Wavelet Toolbox Users
Guide, MathWorks, 1996.
[8] H. Amaris, C. Alvarez, M. Alonso, D. Florez, T. Lobos, P. Janik, J. Rezmer,
Z. Waclawek: Application of advanced signal processing methods for
accurate detection of voltage dips, 13th International Conference on
Harmonics and Quality of Power. ICHQP 2008, Wollongong, Australia,
28th September-1st October 2008, 6p.
[9] CEI EN 50160 Power Quality Application Guide: Voltage Characteristic in
Public Distribution System.

The wavelet approach ensures good performance in presence

of high THD, but the percentage of detected dips is strongly
reduced for high noise level.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
195

Two modified method for Harmonic and Flicker

measurement based on RWPC considering spectral
leakage and edge effects
1

Marjan Saadati , 2Seyed Saeedolah Mortazavi

Shahid Chamran University (SCU)
Golestan Boulevard
Ahvaz, Iran

Abstract- Harmonic is an important issue in electric power

systems. A wavelet packet transform (WPT) method is introduced
as a powerful tool for detection Harmonics and Flicker. The
proposed methods can simultaneously measure the distribution of
RMS quantities with respect to individual frequency bands
directly from the wavelet transform coefficients. Uniform
frequency bands result from the WPT decomposition of power
system waveforms can be used for identification of harmonic
frequency bands. This paper presents two novel modified methods
for harmonic detection and measurement.
However, use of wavelet packet coefficients has some errors such
as the edge effects of wavelet filters and spectral leakage. In first
method paper proposes to combine two adjacent frequency bands
to reduce error caused by spectral leakage. In the recently
methods they place frequency of main harmonics on the edge of
the bands. It provides huge errors because of non-ideal
characteristics of the filters. In second method paper proposes to
change the band width of output to 40 Hz. This causes that the
main harmonics be in the band no on the edge. This method has
got the chance to measure flicker also.
Use of Wavelet Packet Coefficients for harmonic measurement
because of Downsampling reduces number of sampling points and
energy content of signal making huge error. This paper also offers
the use of a compensation method by use of Reconstructed WPC
to prevent energy reduction.
In addition, it introduces a modification method to reduce
edge effects and spectral leakage. Methods are simulated and
experimented. Parameters are compared with true values that it
shows satisfactory results.

I.

INTRODUCTION

POWER quality is becoming an issue of increasing concern

both to utilities and their customers. One of the major power
system problems is steady-state waveform distortion due to
harmonics; Harmonics be produced by variable speed drives,
arc furnaces, personal computers, and other non-linear devices.
Since harmonics can severely degrade the performance of
power system equipment, it is necessary to always monitor
their parameters such as voltage, current, and power.
The traditional discrete Fourier transform (DFT) is proposed

in the and standard as the processing tool for harmonic

analysis, using rectangular time windows of ten cycles width
of the fundamental frequency in a 50-Hz system, providing a
resolution of 5 Hz. Of course this does not preclude the
application of other analysis principles. As is well known, the
results obtained using DFT are incorrect in the case of nonstationary signals. A way to overcome this problem is the use
of the Short Time Fourier Transform (STFT). The STFT
partitions the signal into time segments where the signal is
considered stationary, applying the DFT within each segment.
Once the size of the time window is selected, the timefrequency resolution obtained is fixed and it is the same for the
whole frequency spectrum of the signal. The results obtained
show its dependency on the length of the time window
selected. An advantage of the STFT method is that it gives
information on the magnitude and phase-angle of the
fundamental and harmonics [10].
Kalman filters have been used as an alternative method. This
method gives information both on the magnitude and phase
angle of voltage supply. The detection properties of Kalman
filtering and the accuracy in the estimation depend both on the
model of the system used and on the magnitude, duration and
point-on-wave where the voltage event begins. A possible
solution to improve the performance of Kalman filtering is the
use of an Extended Kalman filter to better estimate the nonlinear process associate with a voltage event.
Wavelet analysis is a powerful signal processing tool
specially useful for the analysis of non-stationary signals. The
discrete wavelet transform provides a non-uniform division of
the time-frequency plane, giving short-time intervals for highfrequency components and long-time intervals for lowfrequency components. The use of wavelets for the analysis of
harmonics has not been thoroughly investigated until now.
Pham and Wong proposed in [2] an approach for identification
of harmonics in power systems using a combination of DWT
and CWT to quantify harmonic frequency amplitudes and
phases. They propose a method to compensate the frequency

Post graduated M.S. degree student in power systems from SCU.

Email address : saadati.s@[Link]
2
Dr. S.S. Mortazavi is Assistant Prof. of Elect Engg of Shahid Chamran University.
Email address : mortazavi_s@[Link]

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
196

response of the filters used in the wavelet transform filter

banks. Hamid and Kawasaki [6] proposed the use of the
wavelet packet transform, with the Vaidyanathan filter, to
improve the results obtained using the discrete wavelet
transform for root mean square values of voltage and power
measurement. Finally Parameswariah and Cox, Eren and
Devaney have studied on the selection of mother wavelet [7].
II.

HARMONIC DETECTION USING WPT

Use of Wavelet Transform as a powerful signal processing

method is receiving increased attention. Initially they use Multi
Resolution Analysis (MRA) for voltage event detection. It
produces nonuniform frequency band that is unsuitable for
harmonic measurements.
To overcome this limitation of the DWT, the wavelet-packet
transform (WPT) can be used to obtain a uniform frequency
decomposition of the input signal as in the Fourier analysis. In
the WPT, both the detail and the approximation coefficients are
decomposed to produce new coefficients, this way enabling a
uniform frequency decomposition of the input signal to be
obtained. By using the WPT and adequately selecting the
sampling frequency and the wavelet decomposition tree, the
output frequency bands of the multiresolution analysis can be
selected to correspond to the frequency bands of the different
harmonic components of the input signal. (The mathematical
background is found in [1]).
III.

HARMONIC DETECTION USING RWPC

However, use of WPC has some errors such as the edge

effects of wavelet filters, the content energy reduction of signal
during decomposition caused by Downsampling and spectral
leakage.
Use of Wavelet Packet Coefficients for harmonic measurement
because of Downsampling reduces number of sampling points
(by power of 2) and energy content of signal making huge
error. This paper proposes using reconstructed wavelet packet
coefficient instead of WPC to prevent energy reduction and
decrease the error. For saving energy coefficients are
reconstructed in each level. Upsampling is used after
downsampling. The other half of the story is how those
components can be assembled back into the original signal
without loss of information. This process is called
reconstruction, or synthesis. The mathematical manipulation
that effects synthesis is called the inverse discrete wavelet
transform (IDWT). Where wavelet analysis involves filtering
and downsampling, the wavelet reconstruction process consists
of upsampling and filtering. Upsampling is the process of
lengthening a signal component by inserting zeros between
samples. The filtering part of the reconstruction process also
bears some discussion, because it is the choice of filters that is
crucial in achieving perfect reconstruction of the original
signal. The low- and highpass decomposition filters (L and H),
together with their associated reconstruction filters (L' and H'),
form a system of what is called quadrature mirror
filters(QMF). Its theory is presented in [9]. Fig. 4 and 5 show
the Reconstruction mechanism and Quadratore Mirror Filters
(QMF) respectively [11].

IV.

MODIFICATION

A signal can be fully decomposed into levels, given by,

where is the total number of data points. Each of these wavelet
levels correspond to a frequency band given by f = 2(fs / N)
Where f is higher frequency limit of the frequency band
represented by the level v; fs is sampling frequency; N is the
number of data points in the original input signal [7].The
maximum frequency that can be measured is given by the
Nyquist theory as fmax= fs/2, Where fs is the sampling
frequency.
V.

METHOD I

The sampling frequency selected is 1.6 kHz and the sampling

window width is 10 cycles of the fundamental frequency (200
ms in a 50-Hz system). The output of the filter bank is divided
into thirty two uniform bands of 25-Hz width (coefficients
d1(n)to d32(n)in Fig. 6). Decomposition is performed in 5 level
and 32 output band. Higher sampling frequencies can be
selected to extend the range of harmonic groups computed in
the input signals. The extension of the sampling frequency
implies the use of a different wavelet decomposition tree to
obtain the same output frequency bands. Thus, doubling the
sampling frequency is needed.
By set frequency band in 25 Hz band width main harmonics set
on the edge of filters. Because of nonideal characteristic of the
filters edge effect error occurs and generates spectral leakage.
This makes measurement so inaccurate specially in frequencies
near fmax/2. An example of frequency response of LP and HP
filters is presented in Fig.7 that shows nonideal characteristics
and edge effect. More investigation of spectral leakage caused
by edge effect is there in [3-5]. Paper has focus on two
important points to generate better accuracy:
1. Choose of mother wavelet. It is important to choose a
wavelet function that has a frequency response near the ideal
and has fast responding. Wavelet mother daubechies with 40
coefficients has good frequency response as shown in Fig.8 [8].
The algorithm uses the DB40 as the wavelet function.
2. Combination of two adjacent frequency bands. The
method proposes to combine two adjacent frequency bands to
reduce error caused by spectral leakage. The outputs of the
filter bank are grouped to produce 15 output bands, with each
harmonic frequency component (both odd and even harmonics)
in the center of each band and with a uniform 50-Hz interval as
depicted in Fig. 9.
VI.

METHOD II

In the recently methods they place frequency of main

harmonics on the edge of the bands [3-5]. It provides some
errors because of nonideal characteristics of the filters. This
makes measurement so inaccurate specially in frequencies near
fmax/[Link] method proposes to change the band width of output
to 40 Hz. This causes that the main harmonics be in the band,
not on the edge. This method has got the chance to measure
flicker also. Then Sampling frequency selected is 2.56 kHz.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
197

Magnitude

Figure [Link] Resolution Analysis (MRA)

Frequency (Hz)
Figure 8. Characteristics of HP& LP filters of DB
Figure [Link] freq. bands of MRA

(a)

(b)
Figure 3.(a) Decomposition tree,(b) Uniform decomposition of WPT

Figure 9. Combination of two adjacent frequency bands that each harmonic

frequency component is in the center of each band and with a uniform 50-Hz
interval.

VII.

Figure [Link] mechanism

Figure [Link] mirror filters (QMF)

Figure 6. Decomposition tree in five levels

The derivation of the rms using discrete wavelet-based

algorithm was proved in [2]. In practice, the analogue
waveforms are digitized. In the WPT algorithm, only the
wavelet coefficients at a certain level j are used for the rms and
power calculations. More information can find in [2].
VIII.

Magnitude

SIMULATION

Consider V(t) as a periodic steady state signal. It has some

harmonics that are mentioned in Table I. This signal is
simulated and by MATLAB and WPT is applied to it. Table I
and II represents the harmonics produced using method I and II
respectively. Figs. 11-14 show the 4 first coefficients that are
obtained. Results show that using RWPC has less error and
better response. In practice signal is produced in 16 cycles and
applied to WPT. The three cycle of beginning and three cycle
of end of the obtained responses are deleted. This is for
elimination edge effect of wavelet filter when applied to signal.
(10 cycles is used.).
IX.

Frequency (Hz)
Figure [Link] characteristics of HP& LP filters

PRESENTATION

CONCLUSION

A wavelet packet transform (WPT) method is introduced as a

powerful tool for detection, classification and quantification of
power quality events. The paper proposed two methods can
simultaneously measure the distribution of RMS quantities
with respect to individual frequency bands directly from the
wavelet transform coefficients. In first method paper proposes

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
198

Magnitude

Magnitude

Frequency (Hz)

Figure [Link] 1
Time

Frequency (Hz)

Magnitude

Figure 12 .coefficient 2

Frequency (Hz)

Magnitude

Figure 13 .coefficient 3

Frequency (Hz)

Figure 14 .coefficient 4

[1]
[2]
[3]
[4]

TABLE II
Simulation results of method II
Bands
1st(flicker)
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
11th
12th
13th
14th
15th
16

Frequenc
y
(Hz)
0-40
40-80
80-120
120-160
160-200
240-280
280-320
320-360
360-400
440-480
480-520
520-560
560-600
600-640
640-680
680-720

True
value
0
230
53
50
44
30
0
6
0
3
0
1
0
0
0
0

Method
using
WPC
0
225.74
51.72
48.29
42.31
26.09
0.26
5.71
0.29
2.82
0.04
1.01
0.09
0.05
0.06
0.08

Modified
method using
RWPC(40)
0.01
227.931
54.68
49.02
41.72
29.11
0.03
5.75
0.31
2.88
0.01
0.99
0
0
0.04
0.01

[5]
[6]
[7]
[8]

Magnitude

to combine two adjacent frequency bands to reduce error

caused by spectral leakage. In second proposes to change the
band width of output to 40 Hz. This causes that the main
harmonics be in the band, not on the edge. This method has got
the chance to measure flicker also. Use of a compensation
method by use of Reconstructed WPC is effective in
prevention in energy reduction. In addition, modification
method to reduce edge effects and spectral leakage is
applicable. Method is simulated and experimented. Parameters
are compared with true values that it shows satisfactory results.

Magnitude

Figure 10. In practice signal is produced in 16 cycles the three cycle of

beginning and end of the obtained responses are deleted for elimination edge
effect of wavelet filter.
TABLE I
Simulation results of method I
True value
Method using WPC
Method using RWPC
Bands
1st
230
225.74
229.91
2nd
53
51.72
52.68
3rd
50
48.29
50.07
4th
44
42.31
43.26
5th
30
26.09
28.79
6th
0
0.26
0.07
7th
6
5.71
6.02
8th
0
0.29
0.31
9th
3
2.82
2.98
th
10
0
0.04
0.02
11th
1
1.01
0.99
12th
0
0.09
0
th
13
0
0.05
0
14th
0
0.06
0.05
15th
0
0.08
0.02

REFERENCES

[Link], M. Prabhakar, Harmonic detection using Wavelet

Transform, IEEE Proceedings of International joint conference on
Neural Networks, Volume 4, 31 July-4 Aug. 2005 Page(s):2228-2231
V.L. Pham and K.P. Wong, Wavelet-transform-based algorithm for
harmonic analysis of power system waveforms, IEE Proc. On
Generation, Transmission and Distribution, 146 (1999) 249-254.
J. Barros and R.I. Diego, Application of the wavelet-packet transform
to the estimation of harmonic groups in current and voltage waveforms,
IEEE Trans. Power Delivery, Vol. 21, No.1, January 2006, pp. 533-535.
Julio Barros , Ramon I. Diego ,A new method for measurement of
harmonic grups in power systems using wavelet analysis in the IEC
standard framework Elsevier power system research. Vol. 76 PP. 200208, 2006
Julio Barros, Ramon I. Diego, Analysis of Harmonics in Power
Systems Using the Wavelet- Packet Transform IEEE Trans.
Instrumentation and Measurement, Vol. 57, No. 1 , Jan 2008
E. Y. Hamid and Z. Kawasaki, Wavelet packet transform for rms
walues and power measurements, IEEE Power Eng. Rev., vol. 21, no.
9, pp. 4951, Sep. 2001.
C. Parameswariah and M. Cox, Frequency characteristics of wavelets,
IEEE Trans. Power Del., vol. 17, no. 3, pp. 800804, Jul. 2002.
A. Mazloomzadeh , M. Mirsalim , H. Fathi and H. Sadeghi, Optimized
method measurement using discrete wavelet packet transform 23th
PSC ,Iran ,1-3 December 2008

[9] Stephane Mallat, A Wavelet Tour of Signal Processing,

Academic Press, London, 2nd Edition, 2001.
[10] Francisco Jurado, Jose R. Saenz , Comparison between
discrete STFT and wavelets for the analysis of power quality
events electric power System Research
[11] MATLAB help on Wavelet.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
199

Combustion Characteristics of Palm Wastes in

Fluidized Bed Combustion
Rosyida Permatasari, Kang Kin Hui, Mohammad Nazri Mohd. Jaafar*
Faculty of Mechanical Engineering, Universiti Teknologi Malaysia, 81310 UTM
Skudai, Johor, Malaysia
prosyida@[Link], nazri@[Link]
Abstract-Fluidized bed combustion (FBC) is one of the most
promising energy conversion options available today. The
emissions from FBC are very dependent on a number of
operating conditions (temperature, staged air, excess air, fuel
feed rate, etc.) and fuel properties. This paper described the
experimental results taken in a staged air fluidized bed
combustion scale laboratory, using palm shell and palm fiber as
fuels and silica sand as the inert bed material. The silica sand
was used for ensuring sustainable fuel ignition and combustion
in FBC. The variation of excess air and fuel feeding rate were
taken in these experiments. Gaseous emissions of CO
concentrations, combustion efficiency and temperature along
the combustor height as well as in flue gas were measured in the
experimental tests. The experimental results showed that the
axial temperature profiles decrease successively related along
the FBC height. The CO emission obtained results lower for
staged air condition than for un-staged air condition. And the
combustion efficiencies give satisfied value. The palm wastes
combustion give significant contribution for reduction of CO
emission from combustion process.
I.

II.

Figure 1. Schematic diagram of the experimental

Combustor

INTRODUCTION

Fluidized Bed Combustion (FBC) is one of the most

promising energy conversion options available today. FBC
uses a continuous stream of air to create turbulence in a
mixed bed of fuel, inert material and coarse fuel ash particles.
It occurs at temperatures typically between 800 and 900C.
Constant mixing of particles encourages rapid heat transfer
and complete combustion. Fluidization, combustion and
emission formation constitute the fundamental issues of FBC.
In order to meet the increasing uses of energy with higher
efficiency without significant environmental impact, the
development and implementation of newer and cleaner
energy conversion system are essential. The increased use of
biomass in energy systems is an important strategy to reduce
the emissions [1]-[3]. Biomass residues from palm oil are
potential usage because of many palm oils planting area in
Asian countries, like in Malaysia, Thailand, and Indonesia
[4]-[5].
The method used in this work was air staging with
controlling of operating conditions, i.e., temperature and
excess air (EA). Air staging was achieved by dividing the air
supply into in-bed fluidizing primary air and over bed
secondary air (SA). Degree of air staging was expressed as
the ratio of secondary to total air. Effects of air staging are
characterized by the reduction of emissions.

EXPERIMENTAL SET-UP

Figure 1 shows a schematic diagram of the experimental

combustor. The experimental combustor was fabricated from
mild steel and is 0.5 m in height and 0.36 m squared in cross
section. The insulated reactor is made from stainless steel
cylindrical tube of 164 mm internal diameter and 2.0 m
height and divided into five flanged sections. Silica sand of
300 m mean particle diameter was used as the bed material.
Air from a blower was introduced into the bed through a
distributor with air outlets arrayed around a circular tube in
six rows. There are a total of 36 air outlets with 6 outlets in
each row of 5 mm diameter each. Reactor preheating was
achieved by using auxiliary fuel. Flue gas exits the top of the
freeboard and enters into a cyclone.

III.

EXPERIMENTAL PROCEDURE

The task of heating the inert material is fulfilled by the preheating system. This was done by introducing the auxiliary
flame directly into the combustion chamber. When the bed
temperature reaches approximately 4500C, fuel was fed by a
screw type feeder at the rate of 79 g/min. Emissions readings
were then taken when the bed temperature stabilizes at
approximately 9500C and above. Air staging is achieved by
dividing the total combustion air supply for stoichiometric

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
200

operation into the primary air (in bed fluidizing air) and
secondary air. Secondary air was introduced at 400 mm
above the distributor plate with a 9.55 mm internal diameter
stainless steel tube. The air staging experiments were
performed with the secondary air to total air ratio varied from
0 to 0.2 at 0.1 increments.

Axial Temperature Profile at 10% Excess Air

Temperature
900.0

800.0

700.0
SA 0%
SA 10%
SA 20%
SA 30%
SA 40%

600.0

TABLE I
FUEL COMPOSITIONS [8]

500.0

Palm Shell

400.0

Type of fuel

Palm Fiber

300.0

Proximate analysis (% by mass)

200

400

600

800

1000

1200

1400

1600

Height above distributor plate (mm)

Moisture
content

12.15

13.98

Ash content

1.96

3.63

Volatile matter

79.22

84.78

Fixed carbon

18.82

11.59

Figure 3. Temperature profile of Palm Fiber for 10% Excess Air

Axial Temperature Profile at 40% Excess Air
Temperature
1000.0
900.0

Ultimate analysis (% by mass)

800.0

Carbon, C

47.978

50.091

Oxygen, O

45.781

41.147

600.0

Hydrogen, H

5.487

6.247

500.0

Nitrogen, N

0.714

2.385

400.0

Sulphur, S

0.04

0.13

300.0

700.0

Other characteristics

SA 0%
SA 10%
SA 20%
SA 30%
SA 40%

200.0
0

Net calorific
value (MJ/kg)
Gross calorific

18.84

17.64

21.44

19.6

200

400

600

800

1000

1200

1400

1600

Height above distributor plate (mm)

Figure 4. Temperature profile of Palm Shell for 40% Excess Air

value (MJ/kg)

Axial Temperature Profile at 40% Excess Air

Temperature
900.0

IV.

RESULT AND DISCUSSION

800.0

4.1. Temperature Distributions

700.0
SA 0%
SA 10%
SA 20%
SA 30%
SA 40%

600.0

Axial Temperature Profile at 10% Excess Air

Temperature
1000.0

500.0

900.0
400.0

800.0
300.0

700.0

SA 0%
SA 10%
SA 20%
SA 30%
SA 40%

600.0
500.0

200

400

600

800

1000

1200

1400

1600

Height above distributor plate (mm)

Figure 5. Temperature profile of Palm Fiber for 40% Excess Air

400.0
300.0
200.0
0

200

400

600

800

1000

1200

1400

1600

Height above distributor plate (mm)

Figure 2. Temperature profile of Palm Shell for 10% Excess Air

Increasing the degree of secondary air tend to decrease

temperature distributions along the FBC height, are showed
in Fig. 2, 3, 4, 5. Increasing the proportion of secondary air in
the combustor reduces flame temperatures and reduced the
residence time due to the local air velocity increase which
increases particles carry-over and volatiles speed.

4.2. CO Emission
Figures 6, 7 show that the CO emission decreases with the
increase of the secondary air to total air ratio. For palm shell
the emission decreases at 30% secondary air (SA) and for
palm fiber the emission decreases and at 20% SA.
Also (has been studied) the effect of excess air (EA) on
formation and reduction of CO in a Fluidized-bed Combustor
Fired with Thai Rice Husk, that CO emission decreases with
increasing excess air.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
201

Effect of Secondary Air Ratio on CO

at Different Excess Air Level

CO index (% vol)

Combustion
efficiency (%)

0.030

Effect of Secondary Air Ratio on Combustion Efficiency

at Different Excess Air Level

95.0
90.0

0.025

85.0

0.020

EA 0%
EA 10%
EA 20%
EA 30%
EA 40%

0.015

0.010

EA 0%
EA 10%
EA 20%
EA 30%
EA 40%

80.0
75.0
70.0
65.0

0.005
60.0

0.000

55.0

10

20

30

40
Secondary air ratio (%)

Fig.6. Effect Secondary Air Ratio on CO Emission of Palm Shell

10

20

30

40
Secondary air ratio (%)

Fig.9. Effect Secondary Air Ratio on Combustion Efficiency of Palm Fiber

Effect of Secondary Air Ratio on CO

at Different Excess Air Level

CO index (% vol)

V. CONCLUSIONS

0.180
0.160
0.140
0.120

EA 0%
EA 10%
EA 20%
EA 30%
EA 40%

0.100
0.080
0.060
0.040
0.020
0.000
0

10

20

30

40
Secondary air ratio (%)

Experiments on the combustion of palm shell and palm

fiber have been successfully carried out in a laboratory scale
fluidized bed combustor using the air staging technique. The
following conclusions can be drawn.
The axial temperature profiles decrease along the FBC
height.
The use of air staging is beneficial for the reduction of CO
emission from a fluidized bed combustor.
The combustion efficiency is also seen to increase with
increased air staging.
The palm wastes combustion give significant contribution for
reduction of CO emission from combustion process.

Fig7. Effect Secondary Air Ratio on CO Emission of Palm Fiber

4.3. Combustion Efficiency

Combustion
efficiency (%)

ACKNOWLEDGMENT
We gratefully acknowledged the financial assistance from
Institutional Grant Universiti Teknologi Malaysia.

Effect of Secondary Air Ratio on Combustion Efficiency

at Different Excess Air Level

65.0
63.0

REFERENCES
61.0
EA 0%
EA 10%
EA 20%
EA 30%
EA 40%

59.0
57.0
55.0
53.0
51.0
49.0
0

10

20

30

40
Secondary air ratio (%)

Fig.8. Effect Secondary Air Ratio on Combustion Efficiency of Palm Shell

From the results of our experiments, it was also observed

that increasing the degree of air staging has the effect of
increasing the combustion efficiency, especially for SA 20%
of EA 20%, 30% and 40%, shown in Figure 8 and 9.
The combustion efficiency maximum for palm shell is 63%
and for palm fiber is 89%.

[1] Wahlund, Bertil. Rational Bioenergy Utilisation in

Energy Systems and Impacts on CO2 Emissions.
Doctoral Thesis. Department of Chemical Engineering
and Technology Energy Processes, Royal Institute of
Technology, 2003.
[2] Nussbaumer T. Combustion and Co-combustion of
Biomass. 12th European Conference and Technology
Exhibition on Biomass for Energy, Industry and Climate
Protection. June 17-12, Amsterdam, 2002.
[3] Slupek S., M. Poskart, L. Szecowka. Reduction of NOx
in Natural Gas Flame using Biomass as Reburning Fuel.
The seventh Asia-PacificInternational Symposium and
energy Utilization, December 15-17, Hong Kong SAR,
2004.
[4] Prasertsan, S. and Prasertsan, P. Biomass Residues from
Palm Oil Mills in Thailand: An Overview on Quantity
and Potential Usage. Biomass andBioenergy, 1996,
11(5), pp 387-395.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
202

[5] Barlow, C., Zen, Z., and Gondowarsito, R. The

Indonesian Oil Palm Industry. International Oil Palm
Study Group, 2003.
[6] Kaewklum, R, et al., Influence of Fuel-moisture
Contentand Excess Air on Formation and Reduction of
CO and NO in a Fluidized-bed Combustor Fired with
Thai Rice Husk, Proceeding of The 2nd Joint International
Conference on Sustainable Energy and Environment (SEE
2006), 21-23 November, Bangkok, Thailand , 2006.

[7] Gibbs, B.M., Pereira, F.J. and Beer, J.M, The Influence
of Air Staging On the NO Emission from a Fluidised
Bed Coal Combustor. Chemical Engineering and Fuel
Technology Department, University of Sheffield,
England,1999
[8] M.R. Taib, L.S. Yang, [Link] Seong, Palm Wastes
Combustion in Bubbling Fluidized Bed. Article,
Perpustakaan Sultanah Zanariah, UTM, Malaysia, 2002.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
203

Dynamic Brake Regulator And Protection System in Locomotive

[Link]
[Link]
Iran university of Science and technology
Email: fa_takbiri@[Link]

Abstract
In this paper, dynamic braking system of
locomotive is simulated by ORCAD
software. Locomotive dynamic braking is a
system which is used to retard locomotive
speed through the conversion of kinetic
energy to electrical energy. This energy
conversion is accomplished by connecting
the traction motors as separately excited
generators with field current being provided
by the main generator. This system consists
of two modules DR and DP. Excitation
current to the motor fields is controlled by
the braking lever position and by the
dynamic braking regulator module DR. The
dynamic brake protection module DP
provides backup protection for the traction
motor fields and the dynamic braking
resistor grids in case a fault develops in the
dynamic braking regulator module DR.
Key words: Dynamic braking system,
Regulator, Traction motor, Resistor grid.

1. Introduction
Locomotive dynamic braking is a system
which is used to retard locomotive speed
through the conversion of kinetic energy to
electrical energy. This energy conversion is
accomplished by connecting the traction
motors as separately excited generators with

field current being provided by the main

generator. The motor armatures are geared

to the axels and rotate whenever the

locomotive is moving. Loading is provided
by connecting the traction motor armature
circuits to dynamic braking grids. Armature
current is determined by the speed at which
the armature rotate and by the amounts of
excitation applied to the motor fields.
With maximum field excitation braking
effort increases from minimum at zero miles
per hour to maximum at approximately 24
miles per hour. Maximum braking effort for
the lower braking lever position is
progressively lower and is attained at
progressively higher track speed as the
braking lever position is decreased. After
maximum braking effort is attained, an
increase in track speed results in a decrease
in braking effort.
The amount of kinetic energy that is
converted into electrical energy is
proportional to
Where is braking grid
current and
is the effective resistance of
the braking grids. The increase in braking
effort from zero to maximum results from
increased motor armature grid current as
track speed increases. This results in an

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
204

BR22

C4
47u

R 18

D 19 D 20 D 21 D 22

2k

C3
47u

6.2 OP2

T12

MR3

PRA2

FTX1

FTX2

EQP2

M F A1

PRA1

M F P11

BWR11 MFP12
M F P13

Z5
10

MR1

BR21
BW A22

Z8 OP1

BW R 13

Z4
10

R24
1.2k

24
26

R6
2.5k
T11

Q
D2

R5
2.5k

BW A21

Q5

RH2
2

22

Q6

1k

BW R 12

M F P21

R23
2.2k

R17
649

Z9
91

R15
5k

D 28

D 11 D 12 D 13 D 14
D 15 D 16 D 17 D 18

LA
ROVER

OP 11

OP1
OP2

Z11
91
Q7

R25

TX2 T21

8
D

3
D

T22
6
D

BW R 23

BWA11
BWA12

Wbreak
BWA2

R10
10k

BWA21
BWA22

Wbreak

R21
50k
Q2
R7

MFP11

1k
Z10
50

1k

MFP1 MFP12
nc
COM MFP13
A no
BRelay
MFP2
+

D1

R2 Z2

R4

32k

1k

BWR1 BWR12
BWR11 nc
COM BWR13
A no
BRelay
BWR2
+

Ground

BWA1

Z1
91

C1
1u

Z6
49.6

Q1

5k

R13
3.65k

28

T21

T12

D 27

R1
2k

R9

R12
8.4k
5

R8
23k

D 10

T22
TX1 T11

R 26

9
D

Q4

RH1
2

R14
2.26k

MFA1
MFA2

Wbreak

BW R 21

To the resistor grids

240k
MFA

20

M F A2

BR13
BR11

P52

M F P23

AL1 AL2
R19
243

R20
550

2. Dynamic brake protection

Q5 is reverse biased during normal

operation. However, if a fault develops in
the dynamic braking regulator module DR,
the excitation voltage applied to the traction
motor fields may tend to rise above a safe
value. Transistor Q5 will become forward
biased if the excitation voltage rises above a
safe value. The collector of Q5 is connected,
through R23, to the base of transistor Q6.
Turn on of Q5 causes Q6 to be forward
biased. Turn on of Q6 causes current to flow
through the LED portion of opto-isolator

+74vdc Ground

1k

R16
7.5

40k

During dynamic braking, zener diode Z8, in

series with resistor R15 and R19, maintain
6.2 volts on the emitter of transistor Q5.
Current flows from terminal 3 of DP module
through the voltage divider consisting of
resistor R19, R16, R17,R18

OVER_EX

Ground
LA RBR

R11

The motor field protection circuit, Fig 1, is

connected across the main generator, in
parallel with the traction motor field, during
dynamic braking. Therefore the motor field
protection circuit detects any changes in
excitation voltage applied to the traction
motor field.

OVER_CU

POS.

NEG.

P51

Excitation current to the motor fields is

controlled by the braking lever position and
by the dynamic braking regulator module
DR. The DR module operates to limit the
excitation current to a value that prevent
armature or grid current from increasing
above the maximum safe current carrying
capacity of the braking grid. The dynamic
brake protection module DP provides
backup protection for the traction motor
fields and the dynamic braking resistor grids
in case a fault develops in the dynamic
braking regulator module DR.

O11. This increases the collector current of

photosensitive transistor portion of O11.
This collector current is applied to the base
of transistor Q7, causing Q7 to be forward
biased. Turn on of Q7 energizes the motor

To the resistor grids

increase of
and consequently an
increase in braking horsepower.

50
R3
2M

C2
33u

MFP21
MFP23

Wbreak

Q3
Q

Z3
91

C5
33u
R22
6.8M

Z7
91
C6
33u

BWR21
BWR23

Wbreak

Fig1. DP module

field protection relay MFP. Pickup of MFP

drops the feed to the equipment protection
relay EQP and recalibrates the motor field
protection circuit by shorting out resistor
R16. Pickup of MFP also provides a positive
feed to the motor field announciator relay
MFA and to the time delay circuit consisting
of R21, R22, C5, and C6.
Drop out EQP drops the feed to the
generator field contactor GFC which
removes excitation voltage from the main
generator field and this decreases the main
generator output voltage. The inductance of
the main generator field windings prevents
an immediate collapse of current through the
field. The decrease in main generator output
Voltage results in a reduction in the voltage

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
205

With forward bias on Q3, the braking lever

signal is removed from the circuit of field.
As a result, excitation to main generator

To the resistor grids

13

field ex.

R4

1.5k

404

925

2
BR121 BR123
D1

D9

T1X
D3

D7

C3
1n

Q
R6
2k

R9
470

Q2

D8

D4

Z1

Z2

Z3

Z4

Z5

Z6

Z7

Z8

Z9

R7

16

16

16

16

16

16

16

16

16

15k

Z14
91
Ground

D2

T1X2

Q1

D 15

R15

R14

96.8

341

D 17
Q4

D5

R13

T2X2

1M
D 16
C6
1n

T2X1
R11
600

D 13

Z13
16

D 25

D 12

Z12
16
R10
7.5k

D 14

Z11
16

D 23

R12
1.5k

C1
1n

D 26

D 24

D 11

Z10
16
BR111

RT4
AL1

field ex.

Q3

R8
360

D6

R5
1.8k

800

BCT

D YN AM I C _BR

R3

SH S22

500

R2

SH S21

BR 131

RH1

SH S12

SH S11

R1

BR 133

RE19

D 22

A signal proportional to braking grid current

is applied to the base of transistor Q1 by a
voltage divider consisting of R5 and R6
connected between the wiper arm of RH1
and the emitter of Q2. When braking grid
current rises above 600 amperes, the voltage
at the wiper arm of RH1 is larger than the
reference signal applied to the emitter of Q2.
This causes current flow from the wiper arm
of RH1 to the zener diodes. This current
flow results in placing forward bias on Q1
and Q2. With forward bias on Q1 and Q2,
current flows through the secondary winding
of transformer T1-A and the primary
winding of transformer T2-A, then from
collector to emitter of Q1 and Q2. This
current induces a voltage into the secondary
T2-A. this voltage is rectified and applied to
a voltage divider consisting of R8 and R9.
Transistor Q3 is forward biased by the
voltage developed across R9.

Output voltage of companion alternator is

applied to the primary of transformer T4 in
series with brake current transductor BCT.
When braking current is low, the reactance
of BCT is high and the voltage applied to
primary of T4 is low. As braking grid
current increases, the output voltage of T4 is
applied between terminal 9 and 11. A
voltage which is proportional to the braking
lever signal, as determined by the braking
lever position, is applied to terminal 7 of DR
module.

D 20

Voltage signals proportional to current

through grids with basic dynamic brakes,
Fig 2, are applied to left port of circuit. The
larger of these signals is applied between
receptacles 1 and 13 of the DR module. The
signal applied between terminal 13 and 1 is
applied to a voltage divider consisting of
rheostat RH1 and resistor R1, R2, R3, and
R4. This same voltage is also applied to
resistor R7 in series with zener diodes Z1
through Z9. The zener diodes provides a
constant reference signal to the emitter of
Q2.

4. Basic dynamic brakes with

trainlined grid current control

D 21

3. Dynamic brake regulator

field decreases. When braking grid current

decreases below 600 amperes, the signal at
wiper arm of RH1 decreases and place a
reverse bias on Q1 and Q2.

D 19

applied to the base of Q5. This reduction in

voltage causes Q5 to become reverse biased.
Reverse bias on Q5 causes MFP to drop out.
Dropout of MFP reestablishes the feed to
EQP and results in reapplying excitation
voltage to main generator field.

BR113 GFC1 GFC2

TX4

10uH

AL2

Fig2. DR module

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
206

Brake current

+74vdc

Transistor Q4 compares the braking lever

signal with the braking grid current signal.
Q4 is forward biased if braking grid current
increases above the braking lever signal.
Turn on of Q4 provides a path for current
flow through transformer T1-B secondary
and T2-B primary. The output of T2-B is
rectified and applied to a voltage divider
consisting of R8 and R9. Forward bias on
Q3 results in decreasing of excitation to the
main generator field.

5. Simulation results
The simulation results applying voltage to
the modules are shown in following figures:

[Link] network current

Fig4. Collector current of Q1

6. References
[1] GT26CW-2 locomotive service manual.
[2] Painter, T.D.; Barkan, C.P.L.; Prospects for
dynamic brake energy recovery on North American
freight locomotives.
4-6 April 2006 Page(s):181 - 188 .IEEE CNF
[3] Grobovoy, A.A.; Denisova, D.V.; Bondareva,
N.V.; Kuznetsova, N.Y.;A fuzzy logic transmission
system dynamic braking control.
27-30 June 2005 Page(s):1 - 7. IEEE CNF
[4] Cochran, Paul;Calculation of Dynamic Braking
Characteristics of Wound Rotor Induction Motors.
Volume 72, Part III, Jan. 1953 Page(s):992 996.
IEEE JNL
[5] Al-Bahrani, A.H.; Malik, N.H.; Selection of the
excitation capacitor for dynamic braking of induction
machines.
Volume 140, Jan. 1993 Page(s):1 6. IET JNL
[6] Joshi, Pradip Raghunath; Dubey, G.
K.;Optimum DC Dynamic Braking Control of an
Induction Motor Using Thyristor Chopper Controlled
Resistance.
Volume IECI-21, May 1974 Page(s):60 65. IEEE
JNL
[7] Rahim, A.H.M.A.; Alamgir, D.A.H.; A closedloop quasi-optical dynamic braking resistor and shunt
reactor control strategy for transient stability [of
synchronous generator].
Volume 3, Aug. 1988 Page(s):879 886. IEEE JNL
[8] Erdman, William L.;Dynamic Braking of DC
Machines: A Mathematical Approach.
Volume IA-19, Part I, May 1983 Page(s):388 392.
IEEE JNL
[9] Hernandez, J.M.; Moreno, R.A.P.;Design of a
variable speed drive with dynamic braking for
induction motor for electric vehicles.
15-19 Oct. 2000 Page(s):211 214. IEEE CNF
[10] Croft, W. H.; Hartley, R. H.; Improving
Transient Stability by Use of Dynamic Braking.
Volume 81, April 1962 Page(s):17 24. IEEE JNL

Fig5. Collector current of Q3

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
207

Lightning Phenomenon Introduction and Basic

Information to Understand the Power of Nature
ukasz Staszewski
University of Technology
Wybrzeze Wyspianskiego 27
Wroclaw, Poland
This paper introduces the history of lightning researches,
some hypothesis of lightning forming process, types of lightning
and triggering ways.

I.

INTRODUCTION

Benjamin Franklin (1706-1790)[1] endeavored to test the

theory that sparks shared some similarity with lightning using a
spire which was being erected in Philadelphia. While waiting
for completion of the spire, he got the idea of instead using a
flying object, such as a kite. During the next thunderstorm,
which was in June 1752, it was reported that he raised a kite,
accompanied by his son as an assistant. On his end of the string
he attached a key, and he tied it to a post with a silk thread. As
time passed, Franklin noticed the loose fibers on the string
stretching out; he then brought his hand close to the key and a
spark jumped the gap. The rain which had fallen during the
storm had soaked the line and made it conductive.
Although experiments from the past time of Benjamin
Franklin showed that lightning was a discharge of static
electricity, there was little improvement in theoretical
understanding of lightning (in particular how it was generated)
for more than 150 years. The impetus for new research came
from the field of power engineering: as power transmission
lines came into service, engineers needed to know much more
about lightning in order to adequately protect lines and
equipment.
Nowadays lightning is considered as an atmospheric
discharge of electricity, which typically occurs during
thunderstorms, and sometimes during volcanic eruptions or
dust storms. In the atmospheric electrical discharge, a leader
of a bolt of lightning can travel at speeds of 60,000 m/s
(220,000 km/h), and can reach temperatures approaching
30,000 C (54,000 F), hot enough to fuse silica sand into
petrified lightning, known scientifically as glass channels[2] or
fulgurites which are normally hollow and can extend some
distance
into
the
ground.
There
are
some
16 million lightning storms in the world every year.
How lightning initially forms is still a matter of debate.
Scientists have studied root causes ranging from atmospheric
perturbations (wind, humidity, friction, and atmospheric
pressure) to the impact of solar wind and accumulation of
charged solar particles. Ice inside a cloud is thought to be a
key element in lightning development, and may cause

a forcible separation of positive and negative charges within

the cloud, thus assisting in the formation of lightning.
II. FORMING PROCESS[3]
1) Charge separation the first process in the generation of
lightning. There are two hypothesis describing the process:
POLARIZATION MECHANISM HYPOTHESIS
The mechanism by which charge separation happens is still
the subject of research, but one hypothesis is the polarization
mechanism, which has two components:
a) Falling droplets of ice and rain become electrically
polarized as they fall through the atmosphere's natural
electric field;
b) Colliding ice particles become charged by
electrostatic induction.
Ice and super-cooled water are the keys to the process.
Violent winds buffet tiny hailstones as they form, causing them
to collide. When the hailstones hit ice crystals, some negative
ions transfer from one particle to another. The smaller, lighter
particles lose negative ions and become positive; the larger,
more massive particles gain negative ions and become negative.

Figure 1. Lighting discharge.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
208

 ELECTROSTATIC INDUCTION HYPOTHESIS

According to the electrostatic induction hypothesis charges
are driven apart by as-yet uncertain processes. Charge
separation appears to require strong updrafts which carry water
droplets upward, super-cooling them to between -10 and 20 C. These collide with ice crystals to form a soft ice-water
mixture called graupel.a The collisions result in a slight
positive charge being transferred to ice crystals, and a slight
negative charge to the graupel. Updrafts drive lighter ice
crystals upwards, causing the cloud top to accumulate
increasing positive charge. The heavier negatively charged
graupel falls towards the middle and lower portions of the
cloud, building up an increasing negative charge. Charge
separation and accumulation continue until the electrical
potential becomes sufficient to initiate lightning discharges,
which occurs when the gathering of positive and negative
charges forms a sufficiently strong electric field.
There are several additional hypotheses for the origin of
charge separation.
According to one such hypothesis, charge separation is
initiated by the ionization of an air molecule by an incoming
cosmic ray.
a
snow pellets
2) Leader Formation
As a thundercloud moves over the Earth's surface, an equal
but opposite charge is induced in the Earth below, and the
induced ground charge follows the movement of the cloud.
An initial bipolar discharge, or path of ionized air, starts
from a negatively charged mixed water and ice region in the
thundercloud. The discharge ionized channels are called
leaders. The negative charged leaders, called a "stepped leader",
proceed generally downward in a number of quick jumps, each
up to 50 meters long. Along the way, the stepped leader may
branch into a number of paths as it continues to descend. The
progression of stepped leaders takes a comparatively long time
(hundreds of milliseconds) to approach the ground. This initial
phase involves a relatively small electric current (tens or
hundreds of amperes), and the leader is almost invisible
compared to the subsequent lightning channel.
When a stepped leader approaches the ground, the presence
of opposite charges on the ground enhances the electric field.
The electric field is highest on trees and tall buildings. If the
electric field is strong enough, a conductive discharge (called a
positive streamer) can develop from these points. This was
first theorized by Heinz Kasemir. As the field increases, the
positive streamer may evolve into a hotter, higher current
leader which eventually connects to the descending stepped
leader from the cloud. It is also possible for many streamers to
develop from many different objects simultaneously, with only
one connecting with the leader and forming the main discharge
path. Photographs have been taken on which non-connected
streamers are clearly visible. When the two leaders meet, the
electric current greatly increases. The region of high current
propagates back up the positive stepped leader into the cloud

with a "return stroke" that is the most luminous part of the

lightning discharge.
3) Discharge
When the electric field becomes strong enough, an electrical
discharge (the bolt of lightning) occurs within clouds or
between clouds and the ground. During the strike, successive
portions of air become a conductive discharge channel as the
electrons and positive ions of air molecules are pulled away
from each other and forced to flow in opposite directions.
The electrical discharge rapidly superheats the discharge
channel, causing the air to expand rapidly and produce a shock
wave heard as thunder. The rolling and gradually dissipating
rumble of thunder is caused by the time delay of sound coming
from different portions of a long stroke.
a) Gurevich's runaway breakdown theory[4]
A theory of lightning initiation, known as the "runaway
breakdown theory", proposed by Aleksandr Gurevich of the
Lebedev Physical Institute in 1992 suggests that lightning
strikes are triggered by cosmic rays which ionize atoms,
releasing electrons that are accelerated by the electric fields,
ionizing other air molecules and making the air conductive by
a runaway breakdown, then "seeding" a lightning strike.
b) Gamma rays and the runaway breakdown theory
It has been discovered in the past 15 years that among the
processes of lightning is some mechanism capable of
generating gamma rays, which escape the atmosphere and are
observed by orbiting spacecraft. Brought to light by NASA's
Gerald Fishman in 1994 in an article in Science, these so-called
Terrestrial Gamma-Ray Flashes (TGFs) were observed by
accident, while he was documenting instances of
extraterrestrial gamma ray bursts observed by the Compton
Gamma Ray Observatory (CGRO). TGFs are much shorter in
duration, however, lasting only about 1 ms.
Professor Umran Inan of Stanford University linked a TGF
to an individual lightning stroke occurring within 1.5 ms of the
TGF event, proving for the first time that the TGF was of
atmospheric origin and associated with lightning strikes.
CGRO recorded only about 77 events in 10 years; however,
more recently the RHESSI spacecraft, as reported by David
Smith of UC Santa Cruz, has been observing TGFs at a much
higher rate, indicating that these occur about 50 times per day
globally (still a very small fraction of the total lightning on the
planet). The voltage levels recorded exceed 20 MeV.
Scientists from Duke University have also been studying the
link between certain lightning events and the mysterious
gamma ray emissions that emanate from the Earth's own
atmosphere, in light of newer observations of TGFs made by
RHESSI. Their study suggests that this gamma radiation
fountains upward from starting points at surprisingly low
altitudes in thunderclouds.
Steven Cummer, from Duke University's Pratt School of
Engineering, said, These are higher energy gamma rays than

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
209

come from the sun. And yet here they are coming from the
kind of terrestrial thunderstorm that we see here all the time.
Early hypotheses of this pointed to lightning generating high
electric fields at altitudes well above the cloud, where the thin
atmosphere allows gamma rays to easily escape into space,
known as relativistic runaway breakdown, similar to the way
sprites are generated. Subsequent evidence has cast doubt,
though, and suggested instead that TGFs may be produced at
the tops of high thunderclouds. Though hindered by
atmospheric absorption of the escaping gamma rays, these
theories do not require the exceptionally high electric fields
that high altitude theories of TGF generation rely on.
The role of TGFs and their relationship to lightning remains
a subject of ongoing scientific study.
4) Re-strike [5]
High speed videos (examined frame-by frame) show that
most lightning strikes are made up of multiple individual
strokes. A typical strike is made of 3 to 4 strokes. There may
be more.
Each re-strike is separated by a relatively large amount of
time, typically 40 to 50 milliseconds. Re-strikes can cause a
noticeable "strobe light" effect.
Each successive stroke is preceded by intermediate dart
leader strokes again to, but weaker than, the initial stepped
leader. The stroke usually re-uses the discharge channel taken
by the previous stroke.
The variations in successive discharges are the result of
smaller regions of charge within the cloud being depleted by
successive strokes.
The sound of thunder from a lightning strike is prolonged by
successive strokes.
III. TYPES OF LIGHTNING
Some lightning strikes take on particular characteristics;
scientists and the public have given names to these various
types of lightning. Most lightning is streak lightning. This is
nothing more than the return stroke, the visible part of the
lightning stroke. Because most of these strokes occur inside a
cloud, we do not see many of the individual return strokes in a
thunderstorm.
The return stroke of a lightning bolt, which is the visible bolt
itself, follows a charge channel only about a half-inch (1.3 cm)
wide. Most lightning bolts are about a mile (1.6 km) long[6].
CLOUD-TO-CLOUD LIGHTNING
Lightning discharges may occur between areas of cloud
having different potentials without contacting the ground.
These are most common between the anvil and lower reaches
of a given thunderstorm. This lightning can sometimes be
observed at great distances at night as so-called "heat
lightning". In such instances, the observer may see only a flash
of light without thunder. The "heat" portion of the term is

Figure 2. Cloud-to-cloud lighting.

a folk association between locally experienced warmth and the

distant lightning flashes.
Another terminology used for cloud-cloud or cloud-cloudground lightning is "Anvil Crawler", due to the habit of the
charge typically originating from beneath or within the anvil
and scrambling through the upper cloud layers of a
thunderstorm, normally generating multiple branch strokes
which are dramatic to witness. These are usually seen as a
thunderstorm passes over you or begins to decay. The most
vivid crawler behavior occurs in well developed thunderstorms
that feature extensive rear anvil shearing.
DRY LIGHTNING[7]
Dry lightning is a term in the United States for lightning that
occurs with no precipitation at the surface. This type of
lightning is the most common natural cause of wildfires.
ROCKET LIGHTNING
It is a form of cloud discharge, generally horizontal and at
cloud base, with a luminous channel appearing to advance
through the air with visually resolvable speed, often
intermittently. It is also one of the rarest of cloud discharges.

Figure 3. Multiple paths cloud-to-cloud lighting.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
210

 GROUND-TO-CLOUD LIGHTNING
Ground-to-cloud lightning is a lightning discharge between
the ground and a cumulonimbus cloud from an upward-moving
leader stroke.
BALL LIGHTNING[8]
Ball lightning is described as a floating, illuminated ball that
occurs during thunderstorms. They can be fast moving, slow
moving or nearly stationary. Some make hissing or crackling
noises or no noise at all. Some have been known to pass
through windows and even dissipate with a bang. Ball
lightning has been described by eyewitnesses but rarely
recorded by meteorologists.
Figure 4. Lightning triggered by a volcanic materials thrust high into
atmosphere.

CLOUD TO GROUND LIGHTNING

Cloud-to-ground lightning is a great lightning discharge
between a cumulonimbus cloud and the ground initiated by the
downward-moving leader stroke. This is the second most
common type of lightning, and poses the greatest threat to life
and property of all known types.
BEAD LIGHTNING
Bead lightning is a type of cloud-to-ground lightning which
appears to break up into a string of short, bright sections, which
last longer than the usual discharge channel. It is fairly rare.
One of theories is the observer sees portions of the lightning
channel end on, and that these portions appear especially bright.
RIBBON LIGHTNING
Ribbon lightning occurs in thunderstorms with high cross
winds and multiple return strokes. The wind will blow each
successive return stroke slightly to one side of the previous
return stroke, causing a ribbon effect.

UPPER-ATMOSPHERIC LIGHTNING[9]
Reports by scientists of strange lightning phenomena above
storms date back to at least 1886. However, it is only in recent
years that fuller investigations have been made. This has
sometimes been called mega-lightning.
IV. TRIGGERING OF LIGHTNING
ROCKET-TRIGGERED
Lightning has been triggered directly by human activity in
several instances. Lightning struck Apollo 12 soon after takeoff.
It has also been triggered by launching lightning rockets
carrying spools of wire into thunderstorms. The wire unwinds
as the rocket ascends, providing a path for lightning. These
bolts are very straight due to the path created by the wire.
VOLCANICALLY TRIGGERED[10]
Extremely large volcanic eruptions, which eject gases and
material high into the atmosphere, can trigger lightning. This
phenomenon was documented by Pliny The Elder during the
AD79 eruption of Vesuvius, in which he perished.
LASER TRIGGERED[11]
Since the 1970s, researchers have attempted to trigger
lightning strikes on demand. Such triggered lightning is
intended to protect rocket launching pads, electric power
facilities, and other sensitive targets.
Researchers generated filamentsb that lived too short a
period to trigger a real lightning strike. Nevertheless, a boost in
electrical activity within the clouds was registered.
b
channel of ionized molecules
REFERENCES
[1]
[2]
[3]
[4]
[5]

Figure 5. Representation of upper-atmospheric lightning and electricaldischarge phenomena.

[Link]
[Link]
[Link]
[Link]
Martin A. Uman (1986). All About Lightning.
Dover Publications, Inc.. pp. 103110. ISBN 0-486-25237-X
[6] [Link]
[7] [Link]
[8] [Link]
[9] [Link]
[10] [Link]
[11] [Link]

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
211

Signal Processing Techniques used in

Power Quality Monitoring
Umar Naseem Khan
from the normal or desired voltage or current sine wave that can
be measured at predefined instances (or more precisely over
predefined windows of time), whereas events are larger disturbances that trigger a recording or further processing [1].
The most common origin of variations are: voltage frequency
variations, voltage magnitude variations; unbalance; voltage
fluctuations or flicker and waveform distortion. While most
common origins of events are interruptions, voltage dips, and
transients.
The issues to be discussed when measuring power quality
variations include: (1) extracting the characteristics; (2) statistics to quantify the performance of the supply at one location;
(3) statistics to quantify the performance of a whole system [1].
The issues to be discussed when measuring events include:
(1) a method has to be defined to obtain the voltage magnitude
from the sampled waveform; (2) threshold levels have to be set
for the beginning threshold and for the ending threshold, these
two thresholds could be the same or different. In addition, a
value has to be chosen for the minimum duration of interruption;
(3) characteristics have to be defined for the event, in this case
the duration of the interruption [1].
A general scheme for carrying out power quality measurements is shown in Fig. 1. The first step in power quality monitoring is the transformation from analog voltages and currents in
the power system to sampled digital values that can be processed automatically. The measurement device block in Fig 1
includes instrument transformers, analog anti-aliasing filters,
sampling and digitizing, and digital anti-aliasing and down
sampling. In this paper, our concentration will be focused on
post processing and we will assume that the sampled and digitized voltage or current waveforms are available for processing.

Abstract This paper gives a general review of different signal

processing techniques which are widely used for the power quality
monitoring. The majority of power quality problems can be
characterized through measurements of voltage and current. To
distinguish the type of disturbances, monitoring systems require
the processing of signals, which concern the extraction of features
and information from measured digital signals. In fact, the use of
signal processing techniques can influence the way that voltage
and current signals are measured and analyzed in power system
field.
Index Termspower quality, signal processing, power quality
monitoring, disturbances, variations, events

I. INTRODUCTION
As it is well known, an ideal three-phase AC supply consists of
three phase voltages that are 120 degrees out of phase and have
identical magnitudes. Above all, these voltages should be sinusoidal and should be available continuously. Any diversion
from these requirements is considered as poor quality of power.
One of the major causes of disturbances in AC power system is
the electronic switching circuits, which are widely used both in
industry and in household systems. Most of the power electronic
equipments are manufactured based on the available standards,
but due to increase in the numbers of such equipments there are
also an increase in disturbances in the power system. While the
sources of disturbances in AC power systems increase continuously, the electric utilities and end users of electrical power
have become more concerned about the quality issues of the AC
power.
The research focuses on power quality monitoring systems are
very much involved in the studies of signal-processing techniques applied to the monitored signals to extract valuable
information for analyzing and diagnosing the cause of power
quality problems. The analysis of this information then help the
power quality experts to determine the sources and causes of
such events and, thus, possible solutions for avoiding these
disturbances can be identified.
The monitoring practice of power quality includes three
steps, detecting and extracting the features of the power quality
events. classifying these events into known waveform categories, and, if the size of the monitored data is very large, compression methods will be necessary for saving these data for
further analysis.

II. DETECTION METHODS OF POWER QUALITY DISTURBANCES

The disturbance capturing techniques are usually based on
detecting the power quality events when a certain threshold
level is exceeded. The limits and threshold levels are primarily
defined by the standards that provide the acceptable level of
disturbance in AC power systems by the end users.

A. Variations and Event

An important division of power quality disturbances is between variations and events. Variations are small deviations
Umar Naseem Khan is with the Faculty of Electrical Engineering, Wroclaw University of Technology, Wroclaw 50-370,
Figure 1. General scheme for power quality measurements
EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
Poland (e-mail:[Link]@[Link]).
212

These methods are divided in two main categories: Detection

methods of variations and detection methods of events. Before,
going further it is important to know about the stationary and
non-stationary signals.
Stationary Signals: A signal is stationary when it is statistical
time invariant (or the statistics of the signal are independent of
time) [1]. All the variations in the voltage and current sine wave
are among the stationary signals.
Non-stationary Signals: If a signal is statistical time varying,
then it is a non-stationary signal.
A. Root Mean Square value (RMS)
The RMS of a signal is not an analysis technique in its own
right, but it gives some basic information about an electrical
system, and is used widely. The main disadvantage of this algorithm is its dependence on the size of the sample window. A
small window makes the RMS parameter less relevant, as it
follows the tendency of the temporal signal, and loses the
meaning of mean value of power. A large window hides the
events [3].
The deviation of the voltage or current signal from the ideal
sine wave is characterized through a number of parameters (or
features): magnitude, frequency, distortion, unbalance, Flicker
Severity and very short variations. We will briefly discuss the
different detection methods for these parameters.

Figure 3. Effect of window size on the rms voltage-time waveform [1]

The effect of changing the length of the window is shown in

Fig 3 for a 1-min window of the phase a voltage of the normal
case. In all four cases, one value per window has been calculated. For a one- or two-cycle window, the rms voltage has a
rather noisy appearance. Applying a longer window gives a
more smooth function of time.
For the three-phase measurement, three methods are mostly
used:
Average of the magnitude values of the three phases:

1) Voltage Magnitude Variations

The vast majority of voltage (magnitude) variation studies
use the rms value to quantify the voltage magnitude. In IEC
61000-4-30, a basic measurement window of 10 or 12 cycles (in
50- and 60-Hz systems, respectively) is prescribed: The measurement window shall be synchronized to the actual power
system frequency. Non-synchronization leads to a small error in
the estimated rms value, where the relative error in rms value is
half the relative error in frequency. A number of alternative
methods are introduced: among others, voltage amplitude,
fundamental component, instantaneous three-phase rms, and
positive-sequence voltage. The latter two values are suitable
alternatives in a three-phase system.
For monitoring purpose, the point-by-point RMS values of
PQ signals have to be calculated, which is defined as
1
nk

V = mean(Va ,Vb ,Vc )

(2)

The rms value of the amplitudes from three phases:

V=

1 2
(Va + Vb 2 + Vc 2 )
3

(3)

The absolute value of the positive-sequence voltage:

V = Vp

(4)

The instantaneous three-phase rms:

V=

1 2
(va (t ) + vb 2 (t ) + vc 2 (t ))
3

(5)

nk + i

v[k ]

2) Voltage Frequency Variations

The standard method to characterize frequency variations consists of counting the voltage zero crossings during a time- period
with an accurately known length (10 s according to IEC
61000-4-30). The high accuracy of this method is based on the
very high accuracy for time measurements. The high accuracy
that can be obtained by the standard method also limits the need
for more advanced methods. Despite this, an alternative method
is introduced, based on the dq-transform that is better suited to
present short-time changes in frequency.
Use of dq-Transform in Three-Phase System
With most power quality measurements, the three phases or line
voltages
are available
forPoland,
processing.
Despite
8. Figure
[Link]
Conference
Engineering,
Karpacz,
May 10-13,
2009 this, the freRMS calculation
of a signalon[4]Environment and Electrical

Vrms ( nk + i) =

(1)

k =i

213

quency is typically only obtained from one channel, the

so-called reference channel. It is however possible to consider
all three phases in the calculation of the frequency by using the
dq-transform, the dq-transform is defined as follows:
vdq (t ) = e j 2 f 0t

1
2 (va (t )ea + vb (t )eb + vc (t )ec )
3

(6)

Where,
ea = 1

Figure 4. Transformation of time signal in its Fourier Transform

1 1
+ j 3
2 2
1 1
ec = a 2 = j 3
2 2
eb = a =

The discrete Fourier transform is thus:

N 1

3) Three-Phase Unbalance
The standard method for quantifying three-phase unbalance is
as the ratio between negative- and positive-sequence voltage.
U

un =

(7)

U+

Alternative definitions for the three-phase unbalance are given

in a number of IEEE standards. When the fundamental components of the phase-to-phase voltages Uab, Ubc, and Uca are
known, the unbalance can be calculated from the expression
1 3 6Q

un =

1 + 1 3Q

U ab 4 + U bc 4 + U ca 4
(U ab 2 + U bc 2 + U ca 2 )2

Under IEC 61000-4-30, the basic measurement window used to

calculate the unbalance is the same as for voltage magnitude and
harmonic distortion: 10 cycles in a 50-Hz system, 12 cycles in a
60-Hz system, about 200 ms in each case. Next to the negative-sequence unbalance, a zero-sequence unbalance may be
calculated as the ratio between zero- and positive-sequence
voltages.
B. Fourier Analysis

One of the most widely used tools in signal processing is Fourier

analysis. This consists of the decomposition of the signal into a
sum of sinusoidal signals of different frequencies. The signal in
the frequency domain is characterized by the angle and the
module of each sinusoidal wave. So this analysis can be viewed
as a mathematical transformation from the time domain to the
frequency domain, as shown in the Fig 4.
The Fourier transform in continuous time of an integrable
function f can be expressed as follows:

f ( ) =

1
2

f ( x )e

xne

2 j
kn
N

n=0

(10)

where
k = 0,....., N 1

The Fourier Transform is very useful in the analysis of harmonics, and is an essential tool for filter design. However, there
are some disadvantages, such as losses of temporal information,
so that it can only be used in the steady state, and cannot show
the moment when the event is produced.
The DFT (Discrete Fourier Transform) has been applied with
great success, thanks to its implementation through its calculus
algorithm, FFT (Fast Fourier Transform), which provides very
efficient and quick calculations.
C. Short Time Fourier Transform analysis

(8)

where
Q=

fk =

j x

dx

(9)

The short-time Fourier transform-m has been used in power

quality analysis due to its applicability to non-stationary signals,
as in the case of most power quality signals [3]. The advantage
of the STFT technique is its ability to give the harmonic content
of the signal at every time-period specified by a defined window. The STFT (VSTFT) of a sampled signal ,v[k], is defined as
[2]:
VSTFT ( , n) =

v[k ].w[k n].e

j k

(11)

Where w = 2n/N, N is the length of v[k]; n= 1,.., N; and w[k- n]

is a selected window that slides over the analyzed signal. The
main disadvantage of this technique is that the greater the
temporal resolution required, the worse the frequency resolution
will be, and vice versa. For a chosen a window size, this will be
equal for all frequencies.
D. Wavelet Transform (WT)

The next logical step is to obtain a window that can vary in time,
so as to be able to get better frequency or time resolution where
necessary[3]. The wavelet transform is based on the ecompostion of a signal in terms of small waves (daughter wavelets)
derived from translation (shifting in time) and dilation (sealing)
of a fixed wavelet function called the "mother wavelet". The

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
214

G. Kalman filters

Figure 5. Analysis of signal with a possible scaled by means of wavelet.

main advantage of the wavelet transform is its varying-length

window. The general formula of the continuous wavelet transform (VWT) of a signal v(k) is given by

Kalman filters are special types of filters. Their solutions are

based on a set of state-space equations. Kalman filters are useful
tools for many power system applications, for example,
real-time tracking harmonics, estimating voltage and current
parameters in power system protection, and estimating the
parameters of transients. Give the observation data z(n), a
Kalman filter is described by a set of state equations and a set of
observation equations as follows[1]:
State equations:
(16)

x(n) = A(n 1) x(n 1) + w(n)

+

v(t ).

VWT (a, b) =

*
a ,b (t ) dt

(12)

(17)

z ( n) = C ( n) x( n) + v( n)

Where a,b(t) is the mother wavelet (the window function, and

the "*" denotes the complex conjugate), and a and b are the
scaling and the translation, respectively [2]. The wavelet allows
the use of long intervals of time where it is necessary a high
resolution in low frequencies and a small intervals where is
required a high frequency information, as it is shown in figure
5[3].
Wavelets have been very useful in electrical transients analysis.
According to the type of disturbance, a different type of wavelet
is used. This renders its application less general.
E. S-Transform (ST)

The S-Transform technique is defined by convolving the analyzed signal, v[k], with a window function. The window function is chosen to be a function of time and frequency, which
gives this technique the advantage over the wavelet technique.
The convolution is performed in the frequency domain by
multiplying the Fourier transforms of the analyzed signal and
the window function. Then taking the inverse Fourier transform
to obtain the analyzed signal in the ST domain. For the discrete
signal with a Gaussian window, the ST of a signal v[k] is calculated as [2]:
n
VST [ k , ] =
N

Observation equations:

N 1

m=0

m+n
V[
].e
N

2 2
n 2 .e j k

where x(n) is a vector of state variables, A(n -1) is the state

transition matrix, w(n) is a vector of model; C(n) connects the
measurement z(n) with the state vector x(n); v(n) is a vector of
observation. For stationary data, A(n) and C(n) are time independent, that is, A(n)= A and C(n) = C.
III. CONCLUSION
Different signal processing techniques has been discussed in
this paper. In terms of calculation RMS techniques is simple,
fast, and much sensitive in sags, swells, and interruptions in the
signals.
Comparison between signal processing techniques
Speed
Sensitivity
Implementation

[1]
[2]

[3]

[4]

The Park's vector approach is based on the locus of the instantaneous spatial vector sum of the three-phase vectors (v1, v2, v3).
The Park's vector component, (vD, vQ) are given by [2]:
2
1
1
v1
v2
v3
3
6
6

VQ =

1
1
v2
v3
2
6

S Transform
Low
High
Difficult

Parks Vector
High
High
Easy

Math H.J Bollen, Irene Yu-Hua Gu, Signal Processing of Power Quality
Disturbances Willy & Sons Publication.
A. M. Gargoom, N. Ertugrul, and W. L. Soong, A Comparative Study on
Effective Signal Processing Tools For Power Quality Monitoring The
University of Adelaide.
T.Y. Vega, V.F. Riog, H.B.S Segundo, Evolution of Signal Processing
Techniques in Power Quality 9th International Conference Electrical
Power, Barcelona 9-11 October 2007
Jurado, F.; Acero, N.; Ogayar, B.; "Application of signal processing tools
for power quality analysis"; Canadian Conference on Electrical and
Computer Engineering, IEEE CCECE 2002. Vol. 1, 12 - 15 May 2002

V. BIOGRAPHY

F. Park's Vector Approach

VD =

Wavelet
Moderate
Moderate
Difficult

IV. REFERENCES

(13)

Where k, m, and n =0, , ,N-1, and V[m+n/N ] is the Fourier

transform of the analyzed signal v[k].

STFT
Moderate
Low
Difficult

(14)
(15)

Umar Naseem Khan was born in 1983 in Pakistan. He

received his [Link]. degree in Electronic Engineering from
Ghulam Ishaq Khan Institute, Pakistan, in 2005. After
getting more than two years of experience in Electric Power
Engineering, he enrolled in [Link]. degree in Control in
Electrical Power Engineering, Wroclaw University of
Technology, Poland, in 2007. Currently with his studies, he
is also attached with R&D Program in Electric Power
Control and Protection with Wroclaw University of Technology and Areva T&D, Poland.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
215

Optimization of the costs of road lighting systems.

Andrzej Makw
Wrocaw University of Technology
Wybrzee Wyspiaskiego 27 50-370
Wrocaw, Poland

I.

INTRODUCTION

Minimization of total costs of construction and exploitation

of road lighting systems becomes a significant topic in Poland
due to high development of road investment acticity in Poland.
Many new motor- and high- ways are built in all regions. They
need to be in use before 2012, because in 2012 we are host of
Euro. Thus high tempo and low costs are recommended.
Optimization of road lighting should include economic aspects
(espenses of lighting systems) carrying on many various
constrains. Optimization of lighting include a lot of tasks. In
paper only most important are consider. More difficult are
neglected and analyses uses a large number of simplification.

where:
H1-height of fitting suspension
H2-height of lighting post
W-bar length
D-distance between the post
Fig.1 Geometry of fittings
A road consists of many various elements.: roadways,
sidewalks, bicycle paths, green belts, crossings etc. Lighting
systems is composed of lighting fittings. Kinds and
configuraton of lighting fittings achieve lighting goal. Lighting
fittings have different sources, rods, posts and equipment. In
paper we consider straight- road case without crossings. There
are many different road lighting configuration: one-sided, twosided opposite, two-sided alternate, on separation belt. Cross
symetry of the fittings is the same as road axis. Big variety of
systems is connected with various road configuration and
lighting requirements. It is main task in construction to choose
right one, which ensures best lighting performance to
considered road. Differenation of the lighting parameters is
achieved by modifying the following features: fitting types,

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
216

way of their arragment, height of fitting suspension, distance

between the posts, projection of the fitting above the road
components, numer of fittings on single post.

II. OPTIMIZATION OF THE ROAD LIGHTING COSTS

The optimization process consists in searching an extreme

value of J(x) function, reffered to as objective function, where
x is for vector of decisive variables. Decisive variables of
objective function are result from many factors as, available
manufacturing technologies and standard requirements. They
can be continous or discrete character. Type and number of
decisive variables influence on optimization method. We can
distinguish two groups of decisive varibles: euality and
inequality character. In the lighting system optimization task
only inequality constrains are dealt. Usually method of penalty
function is used. It consists in replacing the task with
constrains for a constrain-free task by formulation of the
objective function J(x) in modified form J(x):
J/(x)=J(x) + FP(x)

III. CONCLUSIONS

(1)

where J/(x)-modified objective function, FP(x)-penalty

function.
Desisive variables are distance between the fittings, post
height, bar length, fitting upward inclination angle, fitting type,
source type, method of fitting arrangement.
Objective function for road lighting optimization has economic
character. This function is a sum of investment Ji(x) and
operational Je(x) costs.
Ji(x)=KO+ KZ+ Ks+ KOS+Ki+ Kp
Je(x)=Ke+ Kk+ Ks+ Kw+ Kd

The constrains of optimization process of road

lighting system costs can be divided into three groups related
to desisive variables, standard requirment, and the requirments
imposed by investor related (esthetics, shape). The second
group is related to controlling the lighting parameters( average
roadway luminance, total uniformity of lumininace etc.) In
consideration third group is neglected.
Some local extremes may occur during searching
optimal solution to the road lighting system. Vector of decisive
variables x has its constant and discrete components. To solve
the problem of costs optimization of the road lightning system
a nondeterministic method of genetic algorithm is used. A
genetic algorithm is a search technique used in computing to
find exact or approximate solutions to optimization and search
problems. This method has high effectiveness in dealing with
jobs. Disadvantage of method is being frequent solutions that
only approximate the global one. It is necesarry to use some
modifications to improve efficency of method. There are
scaling of the adaption function, use of two-point crossing and
use of tourment selection method with the elements of
exclusivity strategy.

Optimization of a road lighting system consisting in

minimization of total cost is feasible. Genetic algorithm
method should be used for purpose of optimization of systems
with such a objective function and different types of decisive
variables and occurance of local extreme points. Some
modyfication are required to improve quickness of finding an
optimal solution. Decreasing costs od road lighting can rise up
to 20 % in comparision with several constructions.

(2)
(3)

where: Ji(x)-the investment part of objective function, Je(x)operational part of the objective function, KO- the costs of
lighting fittings, KZ- the cost of lighting sources, Ks- the cost of
the posts, KOS- the cost of additional fixtures, Ki- the
installation cost, Kp- cost of design, Ke- the cost of electric
power, Kk- maintenance costs, Kw- the costs of periodical
exchange of the sources, Kd-additional expenses
(transportation,utilization of the sources and fittings).
All solutions of the optimization task, considering all
adopted constrains, is reffered to as an admissible area. The
form of objective function is definite beyond the admissible
area. In case of the jobs of such a type the method of external
penalty function may be used, in the minimization case, for the
points located beyond the admissible area the value of the
objective function is increased. Moreover, the penalty is
different, in accordance with the amount of the transgression
and number of current iteration.

REFERENCES
[1]
[2]
[3]

[Link], [Link], Optimization of the costs of road

lighting,, Oficyna Wydawnicza Politechniki Poznaskiej 2001
[Link], and [Link], Podstawy optymalizacji, 3rd ed.,
Oficyna Wydawnicza Politechniki Warszawskiej 1999.
PN-76/E-02032 Owietlenie drg publicznych,(in Polish)

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
217

Contemporary high voltage gapless surge arresters

Rafal Kurnatowski

Abstract- This paper presents general description of

contemporary high voltage gapless surge arresters. Housing
structure, additional equipment and ZnO element are presented
here.

I.

INTRODUCTION

Idea of using gapless arresters is quite old. First usage of

gapless technology was in 1940. Resistors thyrite were used
then. Their nonlinearity factor was about 3,5. Therefore they
couldnt be used in surge arresters.
Only invention of varistor with nonlinearity factor about 50
in 1968 allowed to get rid of the spark gap. In 1976 first
gapless surge arresters were manufactured.
First generation of varistors were very unstable. Their
leakage current was increasing in time and caused increasing of
temperature and decreasing of energy absorption ability.
Nowadays we have inversed process. Leakage current of
contemporary varistors is decreasing in time which is good.
Structure of varistors should be uniform and isolated from
environment. This and many discovers have been noticed in
nearly 70 years of varistors science

Guaranteed lifetime of porcelain housing is usually equal

50 years. Porcelain surface has hydrophilic properties.
Contemporary scientific research show that breakdown voltage
of porcelain & polymer housing in high humidity conditions
dont differ too much. However in big pollution conditions
polymer insulation electric strength is greater than in porcelain
insulator.

II. TYPES OF HOUSING

First generation of gapless surge arrester had porcelain
housing [1]. Total length of contemporary transmission line
gapless surge arrester 110kV is about 1,3m. To comparison the
length of the same device but with spark gap is equal 2m.
Usually column of varistors is centered in housing but not
always. Usually gas inside is SF6 or nitrogen, sometimes sand.
General division of housing of surge arresters is:
Polymer housing
Porcelain housing
A. Polymer housing or porcelain housing
Basic advantage of polymer materials is hydrophobic
property. The best hydrophobic properties has silicone rubber.
Silicone as outer insulation has been used for over 30 years
with good results. It is true that hydrophobic maintains even if
there is a high pollution level. It is important because
sometimes pollution flashover can occur. Structure of polymer
housed arrester is shown in figure 1.

Figure 1. Structure of polymer housed arrester here ABB design.

III. VARISTORS APPLICATION

A. Structure of ZnO varistor
Zinc oxide ZnO element has been used in gapless high
voltage arresters up to now. Its nonlinearity factor is decreasing
during big surge currents. Practically reduced voltage doesnt
depend on temperature.
These varistors are polycrystalline semiconductors
compound by ZnO grains and other metal oxides (e.g. CoO or
MnO). 97% of all components is ZnO. However contents of
the rest mentioned metal oxides is very important because it

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
218

has big influence to different features of all varistor. Even a

small change in proportions or manufacturing process can
affect the properties of this product. Therefore technological
details of production are a secret of vendors.
TABLE I
MOST OFTEN USED ADDITIVES TO ZINC OXIDE CERAMICS
Chemical element
Co, Mn, Sb
Li

Impact
Create potential barrier on grains border
Reduce leakage current in working voltage conditions

Sb, Si

Reduce grain growth

Be, Ti, Sn

Increase grain growth

Al, Ga, F, Cr
Sb, Ag, V, Ni, Cr
Bi

In before breakdown zone resistance of varistor is of the

order of G and the current in working conditions is of the
order of A. Following equation (1) is empirical formulation
of voltage-current characteristic in breakdown zone.
I = kU

I current, k constant, U voltage, nonlinearity factor.

Contemporary oxide varistors have nonlinearity factor equal
even 80 which depends on manufacture process and on the
current.
IV. ADDITIONAL EQUIPMENT

Increase nonlinearity of U-I characteristic

Stabilization of potential barrier
Creates heterogenic path

Researches of oxide varistors structure were realized

with electron microscopes. Result is visible below at figure 2.
The structure consist grains of ZnO and dissolved cobalt Co
and manganese Mn.

(1)

Additional equipment used with high voltage surge arrester

is: insulted base, line terminals, grounding terminals, surge
counters, function indicators, external grading rings [1].
Stationary 110kV surge arresters are mounted on 1,5m high
reinforced concrete bracket. It protect surge arrester from
intensive snowfall.
To install function indicators bottom of surge arrester cant
be directly grounded. Thats why insulated base is used.
Longer arresters often require external grading rings to
maintain a uniform and acceptable voltage stress along their
length. Operation of such arresters without the grading rings
may lead to failure. We use them for voltages exceed 110kV.
Usually two grading rings are used for 400kV. They are
presented at Figure 4.

Figure 2. ZnO ceramic structure

ZnO grains touch each other without broking of

another phase. Space between these grains is filled with Bi2O3
and spinels Zn7Sb2O12.
B. Varistor conductance phenomena
In varistors voltagecurrent characteristic (Figure 3) we can
distinguish following zones: before breakdown, breakdown and
saturation.

Figure 4. Additional grading rings in surge arrester

The most popular operation indicator is overpressure

indicator. When there is operation the colorful cover is throw
outside the jet.
Counting of operation number is made by special spark gaps.
They also allow to estimate quantity of charge. When spark
gap operate the electric arc leaves at electrodes after-arc track.
Shape of these tracks depend on value and time of arc current.

REFERENCES
[1]
[2]
[3]

[Link], Wysokonapieciowe ograniczniki przepiec, Dolnolskie

Wydawnictwo Edukacyjne, pp. 25-44, Wrocaw 2003.
U. Braunsberger, ZnO arresters in overvoltage protection in pulsed
power circuits, European Electromagnetic Launch Society 12th Topical
Meeting. September 2001, Ayr.
ABB Buyers Guide ,High voltage surge arresters

Figure 3. Voltage-current characteristic of varistor

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
219

Evaluation of Chaotic Ferroresonance in power transformers including Nonlinear

Core Losses
Ataolla Abbasi
Abbasi@[Link]

Mehrdad Rostami
Rostami@[Link]

Hamid Radmanesh
Hamid Reza Abbasi
[Link]@[Link] Abbasi1@[Link]

Faculty of Engineering,
Shahed University,
Khalig-e-Fars free way,
1417953836 Tehran, Iran,
P.O. Box: 18151/159,
Tel: (+98-21)66639821, Fax: (+98-21)51212105, Tehran, Iran

Abstract
This paper investigates the effect of nonlinear core on the
onset of chaotic ferroresonance and duration of transient
chaos in a power transformer. The transformer chosen for
investigation has a rating of 50 MVA, 635.1 kV, the data for
which is given by Dommel etal. (Tutorial Course on Digital
Simulation of Transients in Power Systems (Chapter 14),
IISc, Bangalore, 1983, pp. 17_/38). The magnetization
characteristic of the transformer is modeled by a singlevalue two-term polynomial. The core loss is modeled by a
third order power series in voltage. With nonlinearities in
core loss included, two effects are clear: (i) onset of chaos at
larger values of open phase voltage, (ii) shorter duration of
transient chaos, It also shown that nonlinear core can cause
ferroresonance drop out.

1. Introduction
Ferroresonance is initiated by improper switching operation,
routine switching, or load shedding involving a high voltage
transmission line. It can result in Unpredictable over
voltages and high currents. The prerequisite for
ferroresonance is a circuit containing iron core inductance
and a capacitance. Such a circuit is characterized by
simultaneous existence of several steady-state solutions for a
given set of circuit parameters. The abrupt transition or jump
from one steady state to another is triggered by a
disturbance, switching action or a gradual change in values
of a parameter. Typical cases of ferroresonance are reported
in Refs. [1],[4]. Theory of nonlinear dynamics has been
found to provide deeper insight into the phenomenon. Refs.
[4],[/7] are among the early investigations in applying theory
of bifurcation and chaos to ferroresonance. The
susceptibility of a ferroresonant circuit to a quasiperiodic
and frequency locked oscillations are presented in Refs. [8],[
9], the effect of initial conditions is investigated. Ref. [10]
Isa milestone contribution highlighting the effect of
transformer modeling on the predicted ferroresonance

oscillations. Using a linear model, authors of Ref. [11] have

brought the effect of core loss in damping ferroresonance
oscillations. The importance of treating core loss as a
nonlinear function of voltage is highlighted in Ref. [12]. An
algorithm for calculating core losses from no-load
characteristics is given in Ref. [13].The mitigating effect of
transformer connected in parallel to a MOV arrester is
illustrated in Ref. [14].However, in all the references cited
above, the effect of nonlinear core is either neglected. The
present paper addresses the effect of nonlinear core on the
global behavior of a ferroresonant circuit. The circuit under
study represents a case of ferroresonance that occurred on
1100 kV system of Bonneville Power Administration as
described in Ref.

2. Circuit Descriptions and Modeling

The three-phase diagram for the circuit is shown in Fig. 1.

[Link] Modeling

The 1100 kV transmission line was energized through a

bank of three single-phase as reported in autotransformers.
Ref. [1] ferroresonance occurred in phase A when this phase
was switched off on the low-voltage side of the
autotransformer; phase C was not yet connected to the

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
220

transformer at that time. The autotransformer is modeled by

a T-equivalent circuit with all impedances referred to the
high voltage side. The magnetization branch is modeled by a
nonlinear inductance in parallel with a nonlinear resistance
and
these
represent
the
nonlinear
saturation
characteristic ( i Lm ) and nonlinear hysteresis and eddy
current

characteristics (v m i Rm ) ,

respectively.

The

hysteresis and eddy current characteristics are calculated

from the no-load characteristics by applying the algorithm
given in Ref. [13]. The iron core saturation characteristic is
given by:

i Lm = s 1 + s 2 q (1)

Fig [Link] curve of nonlinear core

The exponent q depends on the degree of saturation. It was

found that for adequate representation of the saturation
characteristics of a power transformer the exponent q may
take the values 5, 7, and 11. In Ref. [15], the core loss is
modeled by a switched resistor; which effectively reduced
the core loss resistance by a factor of four at the time of
onset of ferroresonance. In this paper, the core loss model
adopted is described by a third order power series whose
coefficients are fitted to match the hysteresis and eddy
current nonlinear characteristics given in [1]:
2

i Rm = h 0 + h 1 v m + h 2 v m + h 3 v m

The circuit in Fig. 1 can be reduced to a simple form by

replacing the dotted part with the Thevenin equivalent
circuit as shown in Fig. 4. By using the steady-state solution
of MATLAB Simulink [16] with the data of the 1100 kV
transmission line [1], Eth and Zth were found to be:

(2)

Per unit value of (i Rm ) given in (3)

i ( R) = .000001 + .0047V .0073V 2 + .0039V 3

[Link] circuit of figure 1

Fig. [Link] of Ferro resonance investigations

Fig 2.V-I characteristic of nonlinear core

E th = 130.1kv ; z th = j1.01243E + 0.5

The resulting circuit to be investigated is shown in Fig. 5
where Zth represents the Thevenin impedance. The behavior
of this circuit can be described by the following system of
nonlinear differential equations:

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
221

d
+ a + bq )
dt
2
d
d
L1 ( 2 + a
+ bqq 1 d ) d = 0
dt
dt
1 d
V&c = (
+ a + bq )
c dt

p2 = [e th (t) - v c - p -

E Vc R1 (

R 2 (s1 + s 2 q + h 0 + h1 p + h 2 p 2 + h 3 p 3 )(4)
- L 2 (s1 p + qs 2 (q-1) p ) ]
[L2 (h1 + 2h 2 p + 3h 3 p 2 ) ]

p vc = s1 + s 2 q + h 0 + h1 p + h 2 p 2 + h 3 p 3
C
(5)
With these parameters:
Vbase = 635.1Kv, I base = 78.72 A, Rbase = 8067,
LS p.u = .0188, C p.u = .07955, R p.u = .0014 Rcore = 556.68 p.u

The formulation with ,

p and vc taken as state variables

are given by:

Fig [Link] plan diagram for q=5 with linear Core loss

x1 = ; x2 = p ; x3 = vc

px1 = x 2

(6)

( 7)

px 2 = [e th (t) - x 3 - x 2 q

R 2 (s1 x 1 + s 2 x 1 + h 0 + h 1 x 2 + h 2 x 2 + h 3 x 2 )
- L 2 (s1 x 2 + qs 2 x 1 (q-1) x 2 ) ]
[L 2 (h 1 + 2h 2 x 2 + 3h 3 ?x 2 ?^2 ) ]
(8)

Fig 7. Phase plan diagram for q=7 with linear Core loss

px 3 = s 1 x 1 + s 2 x 1 + h 0 + h 1 x 2 + h 2 x 2 + h 3 x 2 )
C
(9)
For these values of q, system simulated:

for q = 11 s1 = .0667 , s 2 = .0001

for q = 7 s1 = .0067 , s 2 = .001
for q = 5 s1 = .0071 s 2 = .0034
Equations of system with linear core are:

Fig [Link] plane diagram for q=11 with linear Core loss

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
222

3. Simulation Results and Discussion

Time domain simulations were performed using fourth order
Runge_/Kutta method and validated against Matlab
Simulink. The initial conditions as calculated from steadystate solution of Matlab are:

x 1 = 0,0 ; x 2 = 1.67 pu ; x 3 = 1.55 pu)

The circuit in Fig.5 is analyzed by first modeling the core

loss as a constant linear resistance. Figs. 6, 7and Fig. 8 show
the phase plan diagram for q=5, 7 and q=11 with nonlinear
core. Bifurcation diagram for q=/5, 7 and 11 generated by
conventional time domain simulation show in Figs.12, 13
and 14. It was found that the chaotic behavior begins at a
value of ( E p = 0.8 pu ) for q=/5 and ( E p = 1 pu ) for

[Link] plan diagram for q=11 Nonlinear core loss

q=7,11 where represents the amplitude of eth (t).Transient

chaos settling down to the source frequency periodic
solution was observed for some values of parameters as
shown in Fig. 12 for a typical value of E p in the chaotic
Region. Conventional bifurcation diagrams with the
nonlinear model of core loss are presented in Figs. 15, 16
and Fig. 17. Figs.9, 10 and 11 illustrates the corresponding
phase plot from which a period 9 oscillation can be
discerned for q=11.
Fig. 12. Bifurcation diagram for q=5 with linear core loss

Fig 9. Phase plane diagram for q=5 Nonlinear core loss

Fig. 13. Bifurcation diagram for q=7 with linear core loss

Fig10. Phase plane diagram for q=7 Nonlinear core loss

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
223

Fig. 14. Bifurcation diagram for q=11 with linear core loss

loss

Fig.17. Bifurcation diagram for q=11 with nonlinear core

4. Conclusion
The dynamic behavior of a transformer is characterized by
multiple solutions. Inclusion of nonlinearity in the core loss
reveals that solutions are optimistic when compared with
linear models. An operational guideline for the range of
values of the core loss that avoids the jump between steadystate solutions and contains over voltages, is provided.

loss

[Link] diagram for q=5 with nonlinear core

Appendix: Nomenclature

loss

Fig.16. Bifurcation diagram for q=7with nonlinear core

a, b, c index of phase sequence

h0, h1, h2, h3, h4 coefficient for core loss nonlinear function
n
index for the neutral connection
s1
coefficient for linear part of magnetizing curve
s2
coefficient for nonlinear part of magnetizing curve
q
Index of nonlinearity of the magnetizing curve
Zth
Thevenins equivalent impedance
C
linear capacitor
Rm core loss resistance
L
nonlinear magnetizing inductance of the transformer
i
instantaneous value of branch current
v
instantaneous value of the voltage across a branch
element
eth(t ) instantaneous value of Thevenin voltage source
e
instantaneous value of driving source

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
224

p
Eth
Ep
x
f
v

time derivative operator

R.M.S. value of the Thevenin voltage source
peak value of the Thevenin voltage source
state variable
flux linkage in the nonlinear inductance
angular frequency of the driving force

References
[1] H.W. Dommel, A. Yan, R.J.O. De Marcano, A.B.
Miliani, in:H.P. Khincha(Ed.), Tutorial Course on Digital
Simulation of Transients in Power Systems (Chapter 14),
IISc, Bangalore, 1983,pp. 17_/38.
[2] E.J. Dolan, D.A. Gillies, E.W. Kimbark, Ferroresonance
in a transformer switched with an EVH line, IEEE
Transactions on Power Apparatus and Systems PAS-91
(1972) 1273_/1280.
[3] R.P. Aggarwal, M.S. Saxena, B.S. Sharma, S. Kumer, S.
Krishan, Failure of electromagnetic voltage transformer due
to sustained overvoltage on switching*/an in-depth field
investigation and analytical study, IEEE Transactions on
Power Apparatus and Systems PAS-100 (1981) 4448_/4455.
[4] C. Kieny, Application of the bifurcation theory in
studying and understanding the global behavior of a
ferroresonant electric power circuit, IEEE Transactions on
Power Delivery 6 (1991)866_/872.
[5] A.E. Araujo, A.C. Soudack, J.R. Marti, Ferroresonance
in power systems: chaotic behaviour, IEE Proceedings-C
140 (1993) 237_/240.
[6] S. Mozaffari, S. Henschel, A.C. Soudack, Chaotic
ferroresonance
in
power
transformers,
IEE
Proceedings*/Generation Transmission and Distribution 142
(1995) 247_/250.
[7] B.A. Mork, D.L. Stuehm, Application of nonlinear
dynamics and chaos to ferroresonance in distribution
systems, IEEE Transactions on Power Delivery 9 (1994)
1009_/1017.
[8] S.K. Chkravarthy, C.V. Nayar, Frequency-locked and
quasi periodic (QP) oscillations in power systems, IEEE
Transactions on Power Delivery 13 (1997) 560_/569.

[9] S. Mozaffari, M. Sameti, A.C. Soudack, Effect of initial

conditions on chaotic ferroresonance in power transformers,
IEE
Proceedings*/Generation,
Transmission
and
Distribution 144(1997) 456_/460.
[10] B.A. Mork, Five-legged wound*/core transformer
model: derivation, parameters, implementation, and
evaluation, IEEE Transactions on Power Delivery 14 (1999)
1519_/1526.
[11] B.A.T. Al Zahawi, Z. Emin, Y.K. Tong, Chaos in
ferroresonant wound voltage transformers: effect of core
losses
and
universal
circuit
behavioral,
IEE
Proceedings*/Sci. [Link]. 145(1998) 39_/43.
[12] IEEE Working Group on Modeling and Analysis of
Systems Transients, M.R. Iravani, Chair, Modeling and
analysis guidelines for slow transients*/part III: the study of
ferroresonance, IEEE Transactions on Power Delivery, 15
(2000) 255_/265.
[13] W.L.A. Neves, H. Dommel, on modeling iron core
nonlinearities, IEEE Transactions on Power Systems 8
(1993) 417_/425.
[14] K. Al-Anbarri, R. Ramanujam, T. Keerthiga, K.
Kuppusamy,Analysis of nonlinear phenomena in MOV
connected
Transformers,IEE
Proceedings*/Generation
Transmission and Distribution148 (2001) 562_/566.
[15] D.A.N. Jacobson, R.W. Menzies, Investigation of
station service transformer ferroresonance in Manitoba
hydros
230-kV
Dorsey
converter
station,
IPST2001*/International Conference on Power Systems
Transients, Rio de Janeiro (2001). Availablefrom:
[Link]
[16] H.W. Dommel, I.I. Dommel, Transients Program Users
Manual, The University of British Columbia, Vancouver,
1978.
[17] L.O. Chua, P.-M. Lin, Computer-Aided Analysis of
Electronic Circuits (Chapter 17), Prentice-Hall Inc, New
Jersey, 1975, p. 687.
[18] W.C. Rheinboldt, J.V. Burkardt, A locally
parameterized continuation process, ACM Transactions on
Mathematical Software9 (1983) 215_/235.
[19] A. Semlyen, A. Acha, J. Arrillaga, Harmonic Norton
equivalent for the magnetizing branch of a transformer, IEE
Proceedings-C134
(1987)
169.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
225

Primary Energy Market and Electric Market interaction

Nuno Domingues
ISEL, Instituto Superior de Engenharia de Lisboa, Portugal
Rua Conselheiro Emdio Navarro, 1, 1959 007 Lisboa; ndomingues@[Link]

Abstract- A liberalised market is a market where

customers can freely choose their supplier. This
market model came with the introduction of
competition into non-competitive regulated markets.
In a non-competitive environment, supply and prices
are regulated, while in a liberalized market regulation
aims at avoiding the abuse of market power. In the
case of electricity, the starting point was the rigorous
analysis of the sectors of the industry which were
natural monopolies and the identification of the
activities where the barriers to entry were such that no
competition will develop naturally. It had been
assumed in former market organization that the whole
of the electricity industry was a natural monopoly
which led to the acceptance of vertically integrated
monopoly supply utilities. On a closer inspection it
was found that only certain activities were natural
monopolies, in particular the transmission and
distribution networks. This discovery led reformers to
search for mechanisms of introducing competition
wherever this seemed possible. Bearing in mind the
upcoming Iberian Electricity Market MIBEL, we
studied the sensitivity of the Market Clearing Price to
changes in the production inputs, such as coal, natural
gas and fuel.
Index Terms Electricity markets integration,
Sensitivity analysis, Market Clearing Price

1.

INTRODUCTION

arallel to the privatization process of

electricity utilities and the liberalization of
the European electricity markets, the
Portuguese and Spanish markets engaged in a
process of integration which will lead to the
upcoming Iberian Electricity Market (MIBEL).
In both countries it has been introduced changes
in the production, transmission and distribution
of electricity sectors with the aim of achieving a
legally, socially and economically unified
market.

1) The liberalization process in Portugal

The market opening corresponds to the share of
the national electric energy costumers potentially
subjected to the competition and market
mechanisms, integrating the global consumption
of the eligible customers and the share of energy
needs that the entailed deliverers can acquire in
the scope of this free regime. At present, eligible
customers are those at Medium Voltage (MV),
High Voltage (HV) and Very High Voltage
(VHV), with effective annual consumption.
By the end of the 80s it started to be designed
the current model of the Portuguese system
(called SEN Sistema Elctrico Nacional),
which promoted the privatization process
through the publication of the Decree Law (DL)
449/88, of December 10th. The electricity sector
passed through another key moment with the
publication of DL 182 and the 187/95 that
established the organization bases of the SEN
and the principles of the activities of production,
transmission, distribution, cogeneration and
regulation of the Sector. In 19th of December of
1996 the European Union published the
96/92/CE that establishes common rules for the
domestic market of electricity, which had started
in 19th of February of 1997. This European Rule
(CE), negotiated by the State Members of the
European Union, obligated a review in the 1995
legislation, resulting then the DL 56/97. This
also introduced the necessary changes to the
process of privatization of the national electricity
utility (EDP). The law suit was complemented
with the approval of the Proceeding Manual
(MP) by the Regulator (ERSE Entidade
Reguladora dos Servios Energticos), in
29th of February of 2000. So far, the legislative
structure is completed by a set of seven
regulations: four from the ERSE and three from
the Economic Ministry (DGE - Direco Geral
de Energia). [1], [3], [4], [7], [9]

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
226

2) The liberalization process in Spain

In Spain, the government initiated the process in
1998, after the Ministry of the Industry and
Energy signed a protocol with the utilities in
December of 1996. Up to this date the system
was regulated by the MLE (Marco Legal
Estable), which compensated the sunk costs,
aiming the return of investment, the increase of
the operation efficiency and the increase of the
competitiveness of the national production. After
the 70s oil crisis, this body looked for
stimulating the nuclear central offices (a process
that was blocked in 1985 due to fear of terrorist
attacks), the national coal and the lignite (both of
weak quality and low energy rentability that by
itself would not be competitive to the majority
power plants) in order to decrease the existing
dependence to the crude. Since the Spanish civil
war (1936-1939) that the state-owned company
(ENDESA), situated essentially the north and
northeast, turned to these primary energy
sources. Since 1985 the National Grid (REE Red Elctrica de Espaa) operates the net,
dispatch the system and owns the transmission
grid. In 1991 the Iberduero and the
Hidroelctrica Espaola companies joined,
creating the Iberdrola, and in October of 1996
ENDESA (already with the offices Viesgo and
Enher) acquired 75% of the companies FECSA
and Sevillana.
The regulating entity is now the CNE
(Comisin Nacional de Energa) that
replaced the CSEN (Comisin del Sistema
Elctrico Nacional). Since 1996 it strongly
opposed to any company fusion. Due to its
historic agreements, the liberalization process is
difficult and slow. In this scenario, the entrance
of new not competitive producers is possible
which will increase the final tariffs, thus
opposing the initial goals of the free market. In
24th of December of 1997 it was created the
Operadora del
OMEL
(Compaa
Mercado Espaol de Electricidad, S.A.),
related in the article 33 of the Law 54/97 and the
Real Proceeding R.D. 2019/97. [1], [3], [4], [6],
[7]
2.

DATA COLLECTION

For the purposes of our study I started by

searching the market prices for different inputs:
coal, fuel, natural gas and diesel. Based on
different load factors of the different

technologies and different installed capacities,

applying a conversion factor (from thermal to
electrical energy) and the plant efficiency, it was
possible to reach the marginal cost of production
for each technology. Then, a quadratic cost
function was estimated for the coal, fuel, natural
gas and diesel power plants, applying a different
approach to the hydro, nuclear and renewable
production. Concerning the hydro power plants it
was considered its installed capacity. It was
considered, thus, an almost constant and coherent
cost with the remaining costs. For the nuclear,
constant cost for all was considered. For the
renewals, as they are not dispatchable, as
established in D.L. 339-C/2001, its production
was accounted as a load decrease. In parallel, the
characterization of the installed power and the
produced energy of the Portuguese and Spanish
grids were made. One can, then, shape the curves
of cost of production curve of the Portuguese,
Spanish and the aggregate System. [4], [5], [7],
[8]
3.

SIMULATION RESULTS

The simulation results for the winter scenario

(PH) are:
- Market-clearing price (MCP):
32.4075 /
MWh
- Market-clearing quantity (MCQ): 46165.7028
MWh
And for the summer scenario (VS) are:
- Market-clearing price (MCP): 11.7925 / MWh
- Market-clearing quantity (MCQ): 22420.0432
MWh
Changing the price of the fuel in percentage, I
achieved the values for the MCP presented in
Figure 1 in the appendix.
For each scenario I have represented the
corresponding trend line, as shown in fig. 2 and
in fig. 3 in the appendix.
In the winter scenario (PH) the water resources
are not enough to satisfy the demand. The
cheapest fuel is the coal, so will be the next
technology to be used. The MCP is immune to
coal price variations indicating that the marginal
production came from another technology.
Follows the electric production by oil resource. I
can see that there are some competition between
oil and the gas. So, if the price of any of these
two change there is a reflection in the MCP. If
the price of oil decreases there is no change in

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
227

MCP because the marginal production

technology is gas. However, as the price of oil
raises, less units will be produced by this
technology and the MCP will also raise because
there will be competition between gas and oil. In
the Summer scenario (VS) the water and the coal
fuels are enough to satisfy the demand. So,
changes in gas and oil prices will not be reflected
in MCP because neither are the marginal
technology. However, raising the price of the
coal too much or decreasing the price of oil too
much will bring some competition between them
and will be reflected in MCP.
Note that MCP goes from 32 /MWh to
11 /MWh as the MWh units produced goes
from 46 165 to 22 420.
4.

CONCLUSIONS

It is often being argued that a liberalised market

will lead to real competition, with the
consequence of welfare improvements, but that
is not necessarily true. The basic assumption of a
competitive market is, among others, the
possibility to choose between different suppliers
and that these suppliers are able to generate
electricity at the lowest possible costs.
This paper focus on the sensitivity of the Market
Clearing Price (MCP) to the change of the fuel
price. Other sensitivities can be explored, such as
demand curve, lost of an interconnection, lack of
water resources, among others. One of the
powerful tools of this simulator is analyse these
kinds of sensitivities. Others are the impacts of

6.

the aggregation for both Countries and for all

players.
I stress that in a liberalized market volatility of
MCP is a key issue. Demand forecast is quite
difficult if compared to other commodities,
giving a wider confidence range. Risk
management joins this two properties and faces
possible errors in the forecasting tools.
5.

REFERENCES

[1] Designing competitive electricity markets;

1997 Hung po Chao, Hillard [Link];
Kluwer Academic Publisher
[2] Electric power distribution reliability; 2002
Richard E. Brown; ABB inc. Raleigh North
Carolina
[3] Electricity Market Reform, 2002 IEA
(International Energy Agency) HandBook
[4] Restructured Electrical Power Systems
Operation, Trading and Volatility; Mohammad
Shahidehpour e Muwaffaq Alomoush; Marcel
Dekker, Inc; 2002;
[5] Market Operations in Electrical Power
Systems Forecasting, Scheduling and Risk
Management; Mohammad Shahidehpour, Hatim
Yamim e Zuyi Li; A John Wiley & Sons inc;
2001;
[6] Deregulation on Electric Markets,
Comunicao Introducing Competition to
electricity industry in Spain, 2000
[7] Electric Power Systems, B.M. Weedy, B.J.
Cory, 4 edio, 1998
[8] Cogeneration, Operation and Control,
Welley, 1996
[9] [Link]

APPENDIX

Fuel Price
Technology
Coal

Oil

Gas

Scenarios

80,00%

90,00%

100,00%

110,00%

120,00%

PH

32,41

32,41

32,41

32,41

32,41

VS

9,43

10,61

11,79

12,97

14,15

PH

32,40

32,40

32,41

33,10

33,41

VS

11,79

11,79

11,79

11,79

11,79

PH

26,73

29,87

32,41

35,64

38,88

VS

11,79

11,79

11,79

11,79

11,79

Figure 1: MCP sensitive table for fuel price changes in /MWh

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
228

Winte r

Market Clearing Price (/MWh)

39,00
37,00
35,00
Coal

33,00

Oil
31,00

Gas

29,00
27,00
25,00
70%

80%

90%

100%

110%

120%

130%

Fue l Price

Figure 2: MCP sensitive table for fuel price changes in the Winter scenario

Market Clearing Price (/MWh)

Summer
15,00
14,00
13,00
Coal
12,00

Oil
Gas

11,00
10,00
9,00
70,00%

80,00%

90,00% 100,00% 110,00% 120,00% 130,00%

Fuel Price

Figure 3: MCP sensitive table for fuel price changes in the Summer scenario

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
229

Measurements of electromagnetic compatibility using

GTEM chambers
Marcin Dbowski

Wroclaw University of Technology

Wrocaw, Poland

Abstract The paper will give basic knowledge about EMC

from measurements point of view. Principals of tests done in
GTEM chamber will be presented.

I.

INTRODUCTION

Electromagnetic compatibility is relatively new concept. The

producers of devices sold in the European Union common
market are obligated to check EMC from year 1996. EMC is
nothing else then ability of the device, installation or system to
proper work in electromagnetic environment without adding
additional distortions to this environment or other working
devices. Because there are more then one possible scenario of
interferences between two devices, the main problem if how to
analysis those distortions. For this goal the GTEM chamber
was design and build.
II. THEORY
From the entire EMC test, the measurements of radiation
emission are the hardest, most time consuming and most
expensive, because they need special, big enough field with
low external radiation emission coming from other sources.
Most of todays directives concerning EMC test, recommend
performing the measurements in open area test site (OATS).
Main elements of this environment are:
- low ambient and noise floor,
- oversized ground plane,
- host equipment area,
- Weather protection.
Building such location is very hard due to weather
conditions and electromagnetic wave propagation. Taking into

account many inconveniences of this method, other one is

needed, which will allow carrying out test in more friendly
conditions. One of the most common techniques is the
anechoic and shielded chamber, which are independent form
external weather and electromagnetic conditions.
III. CONSTRUCTON OF GTEM CHAMBER
The GTEM cell (Gigahertz Transverse Electromagnetic cell)
is usually a construction in shape of the rectangular prism,
equipped with screens limiting influence of external
electromagnetic fields and special internal shields preventing
leakage of electromagnetic wave generated in the chamber.
With proper shielding level and wave absorption (absorbing
function is carried out be special absorbers build in the
chamber) it is possible to obtain restricted measurement area,
which is usually few times smaller then dimensions of the
chamber. In this area propagation of electromagnetic wave is
know and can be controlled.
The GTEM chamber are developed version of TEM chamber,
they are able to measure frequencies up to several GHz.
IV. MEASUREMENTS
The GTEM chamber is used to measure intensity of field
distortions produced by examined devices. Basic equipment is
the chamber and measurement device.
Emission of investigated device is obtained by measuring
voltage at the output of the chamber for 3 planes of examined
object. The emission distortion level is calculated basing on the
dependence between obtained values and field intensity values

Absorber

Point of field
calibration
Work space

Figure 1. Schema of GTEM chamber

Figure 2. GTEM chamber in Laboratorium Urzdze Elektronicznych
Pozna

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
230

V. SUMMARY

measured on OATS.
To fulfill standard EN 55022 the analyzer must be set to
check the object in the range of frequencies from 30 MHz to 1
GHz with step of 120 Hz. The detector must be able to measure
quasi peak value for which maximum emission levels are
defined table 1. Other values like peak, average, RMS are
optional.

Recently electronic and electric devices were checked only

in a meaning of fundamental requirements. The EMC tests
were not so important to producer. Now all producers must
fulfill conditions given by EN 55022, because of that they are
obligated to create the technical documentation of the product.
The documentation will give overview if the object is safe in
meaning of EMC. What is more the manufacturer need to
The main problem with performing the measurements is the guarantee the production procedure according to created
time. It is easy to calculate that 8084 measures need to be taken. documents.
The quasi peak detector needs at least 1 second to register the
It is estimated that 10% of nowadays devices are connected
value. For 1 plane it is more then 2 hours. This time needs to with usage of components that will not cause problem in point
by multiply be 3, because the tests are performed in 3 planes.
of electromagnetic compatibility.
To circumvent this barrier special procedure was created:
Good strategy is to test main elements of designed device
- pre-testes for whole frequencies range with peak value separately it will give lower probability that they will cause
detector,
problem working together.
- choice of frequencies were the intensity exceeds or is
REFERENCES
close to threshold,
-

precise tests for chosen frequencies with quasi-peak

detector.
All three points are performed for each plane.
Table 1
Permissible values of quasi-peak value for IT devices according to EN 55022
Permissible level [dBuV/m]
Frequencies range
Class A devices
Class B devices
[MHz]
Quasi-peak value
Quasi-peak value
30 230
40
30
230 1000
47
37

[1]
[2]
[3]
[4]
[5]

Ph. D. Maik Honscha Lecture material for Introduction to EMC BTU

Cottbus 2008.
Jarosaw uszcz, Artur Knitter Badanie emisyjnoci promieniowanej
urzdze energoelektronicznych w komorach GTEM, XV Seminarium
zastosowanie komputerw w nauce i technice 2005.
Clemens Icheln The construction and application of a GTEM cell, HUT
Hmaburg 1995
Jan Bogucki, Andrzej Chudziski, Justyn Poujan Emisja
elektromagnetyczna urzdze w praktyce Telekomunikacja i techniki
informacyjne 1-2/2007.
Krzysztof Sieczkarek Badania Kompatybilnoci elektromagnetycznej
Instytut Logistyki i Magazynowania Pozna

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
231

Distributed Generation and Power Quality

Umar Naseem Khan
[2]. Generators larger than this are typically interconnected at
transmission voltages where the system is designed to accommodate many generators. The normal distribution system delivers electric energy through wires from a single source of
power to a multitude of loads. Thus, several power quality
issues arise when there are multiple sources.

Abstract The main purpose of this paper is to discuss the basic

understanding of power quality in relation to the distributed generation. Due to considerable overlap between two technologies,
disturbances affecting the power quality, which are mainly cause
by the addition of Distributed Generation (DG) on the existing
power system network. Injection of the DG into an electric power
grid can affect the voltage quality. Distributed generation of different voltage levels when connected to the power system network
could influence the voltage regulation, sustained interruptions,
harmonics, sags, swells, etc. All the information given here is collected from different references by keeping in mind the students at
the beginning level of the concerned topic.

II. INTERFACE TO THE UTILITY

Here we are only concerned about the impact of distributed
generation on power quality. While the energy conversion
technology may play some role in the power quality, most
power quality issues relate to the type of electrical system interface. Some notable exceptions include: (a) The power variation from renewable sources such as wind and solar can cause
voltage fluctuations. (b) Some fuel cells and micro turbines do
not follow step changes in load well and must be supplemented
with battery or flywheel storage to achieve the improved reliability expected from standby power applications. (c) Misfiring
of reciprocating engines can lead to a persistent and irritating
type of flicker, particularly if it is magnified by the response of
the power system.
The main types of electrical system interfaces are synchronous
machines, asynchronous (induction) machines and electronic
power inverters [2].
Synchronous Machines: Some actual examples of unexpected consequences are
1. The harmonic voltage distortion increases to intolerable
levels when the generator is attempting to supply adjustable-speed-drive loads.
2. There is not enough fault current to trip breakers or blow
fuses that were sized based on the power system contribution.
3. The voltage sag when elevator motors are being started
causes fluorescent lamps to extinguish.
Generators must be sized considerably larger than the load to
achieve satisfactory power quality in isolated operation.
Asynchronous (induction) machines: Induction generators
are induction motors that are driven slightly faster than synchronous speed. They require another source to provide excitation. The requirements for operating an induction generator
are essentially the same as for operating an induction motor of
the same size. The chief issue is that a simple induction generator requires reactive power (vars) to excite the machine from
the power system to which it is connected. To supply the reactive power locally, power factor correction capacitors are
added. While this works well most of the time, it can bring about
another set of power quality problems. One of the problems is
that the capacitor bank will yield resonances that coincide with
harmonics produced in the same facility. Another issue is
self-excitation. An induction generator that is suddenly isolated
on a capacitor bank can continue to generate for some period of
time. This is an unregulated voltage and will likely deviate
outside the normal range quickly and be detected.

Index Termspower quality, distributed generation, distributed generators, distributed resources, disturbances.

I. INTRODUCTION
The demand of power is escalating in the world of electricity.
This growth of demand triggers a need of more power generation. DG uses smaller-sized generators than does the typical
central station plant. Distributed generators are small scale
generators located close to consumers; normally Distributed
Generators are of 1 kW to 100 MW [1].
Definition of DG [2]
Distributed generation in simple term can be defined as a
small-scale generation. It is active power generating unit that is
connected at distribution level.
IEEE defines the generation of electricity by facilities sufficiently smaller than central plants, usually 10 MW or less,
so as to allow interconnection at nearly any point in the
power system, as Distributed Resources.
Electric Power Research Institute (EPRI) defines distributed
generation as generation from a few kilowatts up to 50 MW.
International Energy Agency (IEA) defines DG as Power
generation equipment and system used generally at distribution levels and where the power is mainly used locally on
site.
The International Council on Large Electricity Systems
(CIGRE) defines DG as generation that is not centrally
planned, centrally dispatched at present, usually connected
to the distribution network, and smaller than 50-100 MW.
These generators are distributed throughout the power system
closer to the loads. The DG penetration in the grid poses new
challenges and problems to the network operators as these can
have a significant impact on the system and equipment operations in terms of steady-state operation, dynamic operation,
reliability, power quality, stability and safety for both customers
and electricity suppliers. However as we are only concerned
with power quality of the primary and secondary distribution
system, we will only consider generator sizes less than 10MW

Umar Naseem Khan is with the Faculty of Electrical Engineering, Wroclaw University of Technology, Wroclaw 50-370,
EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
Poland (e-mail:8. [Link]@[Link]).
232

Electronic power inverters: All DG technologies that generate either dc or nonpower frequency ac must use an electronic power inverter to interface with the electrical. The early
thyristor-based, line-commutated inverters quickly developed a
reputation for being undesirable on the power system. The
line-commutated inverters produce harmonic currents in similar
proportion to loads with traditional thyristor-based converters.
Besides contributing to the distortion on the feeders, one fear
was that this type of DG would produce a significant amount of
power at the harmonic frequencies. Such power does little more
than heat up wires. To achieve better control and to avoid
harmonics problems, the inverter technology has changed to
switched, pulse-width modulated technologies [2].
III. POWER QUALITY ISSUES
A major issue related to interconnection of distributed resources onto the power grid is the potential impacts on the
quality of power provided to other customers connected to the
grid.
A. Voltage Regulation

Over-voltages due to reverse power flow: If the downstream

DG output exceeds the downstream feeder load, there is an
increase in feeder voltage with increasing distance. If the substation end voltage is held to near the maximum allowable
value, voltages downstream on the feeder can exceed the acceptable range.
Interaction with load tap changers (LTC) and static voltage
regulators (SVR) controls: The presence of DG can cause localized changes in flow patterns, which are not reflective of the
general trend on the feeder. As a result, LTC or SVR can be set
such that a good voltage profile may not be obtained.
Figure 1 illustrates one voltage regulation problem that can
arise when the total DG capacity on a feeder becomes significant. This problem is a consequence of the requirement to
disconnect all DG when a fault occurs.
Fig 1a shows the voltage profile along the feeder prior to the
fault occurring. The intent of the voltage regulation scheme is to
keep the voltage magnitude between the two limits shown. In
this case, the DG helps keep the voltage above the minimum
and, in fact, is large enough to give a slight voltage rise toward
the end of the feeder.

When the fault occurs, the DG disconnects and may remain

disconnected for up to 5 min. The breaker recloses within a few
seconds, resulting in the condition shown in Fig. 1b. The load is
now too great for the feeder and the present settings of the
voltage regulation devices.
Therefore, the voltage at the end of the feeder sags below the
minimum and will remain low until voltage regulation equipment can react. This can be the better part of a minute or longer,
which increases the risk of damage to load equipment due to
excessively low voltages.
Solutions include:
1. Requiring customer load to disconnect with the DG. This
may not be practical for widespread residential and small
commercial loads. Also, it is difficult to make this transition
seamlessly and the load may suffer downtime anyway, negating
positive reliability benefits of DG.
2. Installing more voltage regulators, each with the ability to
bypass the normal time delay of 30 to 45 s and begin changing
taps immediately. This will minimize the inconvenience to other
customers.
3. Allow DG to reconnect more quickly than the standard
5-min disconnect time. This would be done more safely by using
direct communications between the DG and utility system
control.
4. Limit the amount of DG on the feeder.
B. DG Grounding Issue:

A grid-connected DG, whether directly or through a transformer, should provide an effective ground to prevent
un-faulted phases from over-voltage during a single-phase to
ground fault. Proper grounding analysis of DG will ensure
compatibility with grounding for both the primary and secondary power systems. This analysis must consider (1) the generator-winding configuration (or inverter arrangement), (2) its
grounding point, (3) the interface transformer configuration,
and (4) grounding of both the primary and secondary power
systems to which the DR is connected [4].
Solutions Include[1]:
Grounding recommendations for distributed resources:
Design of Primary
Distributed System

Three-wire, ungrounded system or

high-impedance,
grounded system

Design of Secondary System

4-wire, Grounded

3-wire, ungrounded

4-wire, Grounded
Four-wire,
multi-grounded
neutral system

3-wire, ungrounded

DR Grounding
DR should be ungrounded or
high-impedance grounded with
respect to primary and effectively grounded with respect to
the secondary system
DR should be ungrounded or
high-impedance grounded with
respect to primary and secondary system
DR should be effectively
grounded with respect to the
primary and secondary system.
Dr should be effectively
grounded with respect to the
primary system and ungrounded or high-impedance
grounded with respect to the
secondary system

EEEIC
International
onclear
Environment
and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
Figure 1. Voltage 8.
profile
change
when DG Conference
is forced off to
faults.
233

C. Harmonic Distortion

E. Islanding

Voltage harmonics are virtually always present on the utility

grid. Nonlinear loads, power electronic loads, and rectifiers and
inverters in motor drives are some sources that produce harmonics. The effects of the harmonics include overheating and
equipment failure, faulty operation of protective devices, nuisance tripping of a sensitive load and interference with communication circuits.
All power electronic equipments create current distortion that
can impact neighboring equipment. DG like PV, fuel cells are
likely to introduce harmonics problem in the system.
Harmonics from DG come from inverters and some synchronous machines. The PWM (pulse width modulation) switching
inverters produce a much lower harmonic current content than
earlier line-commutated, thyristor-based inverters [1].
One new distortion problem that arises with the modern inverters is that the switching frequencies will occasionally excite
resonances in the primary distribution system. This creates
non-harmonic frequency signals typically at the 35th harmonic
and higher riding on the voltage waveform. This has an impact
on clocks and other circuitry that depend on a clean voltage zero
crossing. A typical situation in which this might occur is an
industrial park fed by its own substation and containing a few
thousand feet of cable. A quick fix is to add more capacitance in
the form of power factor correction capacitors, being careful not
to cause additional harmful resonances [1].
Solutions include:
1. Newer PWM inverters have lower current distortion
2. Use non-resonant switching frequencies
3. Use reactors in the neutral, or generators with a 2/3 coil
winding pitch

Refers to a condition in which distributed generation is isolated on a portion of the load served by the utility power system.
It is usually an undesirable situation, although there are situations where controlled islands can improve the system reliability. Islands may be intentional or unintentional [2].
If an island should occur, it should persist for only a very
brief period, unless the aggregate real and reactive output of all
the DG supporting the island is close to the load demand. Otherwise, island voltage and frequency will change rapidly and all
the DG has to be shut down to prevent this.
In case the DG in the distribution system is capable to meet
the load demand, DG can be operated in the island mode and
continue to energize the distribution system. But the major
issues with this type of inadvertent islanding are:
1.
The voltage and frequency provided to other customers
connected to the island are out of the utilitys control,
yet the utility remains responsible to those customers.
2.
Protection systems on the island are likely to be uncoordinated, due to the drastic change in short-circuit
current availability.
Out-of-step reclosing: Many utilities use an instantaneous
reclosing practice, where breakers and circuit reclosers reenergize the protected circuit without any intentional delay and this
could result in out of phase reclosing of the distribution system.
As a result of out of phase reclosing:
1. Large mechanical torques and currents are created,
which can damage the generator or the prime mover.
2. Transients are created which are potentially damaging
to utility and other customer equipment.
3. Out-of-phase reclosing, if it occurs at a voltage peak,
will generate a very severe capacitive switching transient. In a lightly damped system, the crest over-voltage
can approach three times rated voltage.
Prevention [1]:
1. Inverter controls are designed to raise a rising frequency
or lower a dropping frequency
2. The power system frequency acts to correct the inverter
frequency
3. Without the power system to correct the frequency, the
destabilizing signal in the inverter control quickly causes
an over- or under-frequency condition, and frequency
relays trip the inverter
4. Load/generation imbalance relies on an intentional and
significant difference between the DG output and the
local load. DG is operated at constant power factor or
constant reactive power, and not permitted to regulate
voltage. When an island forms, the mismatch between
the DG and the load will quickly cause detectable voltage and/or frequency variations

D. Flicker

Some energy source (e.g., wind turbine or fuel cell) has some
mechanical (or chemical) fluctuations in power output and some
electrical equipment (e.g., the dc bus and inverter) does not have
sufficient energy storage to smooth out these fluctuations. This
will result in fluctuations in the power delivered by a DG and
can cause flicker in the power system in a fashion very similar to
that caused by load fluctuations [3].
Solutions include [3]:
1. Utility companies try to limit flicker so that it is at a level
that cannot be perceived by the human eye. This is accomplished by designing the power system to be sufficiently robust so that smaller load variations do not
create noticeable voltage variations.
2. It is also controlled by imposing limits on the types of
loads that are allowed to connect at various points on the
system
3. When a larger DR unit is applied on a feeder, rapid
response voltage regulators (static VAR compensators)
or fast-response reactive compensation using inverter
reactive-power capabilities can do mitigation of flicker.
4. Energy storage technologies can be applied to smooth
the output fluctuations of solar and wind energy systems.

F. Protection System [4]

Tradition distribution systems were not designed to have active power generating units in them. Power is supplied by the
transmission system and power flow is mainly unidirectional.
But with the DG in the system, power flow can be bi-directional.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
234

8. Unsynchronized reclosing
Solutions Include
1. Reduction of Reach: Adjust relay to increase reach.
Add recloser to add another protection zone. Minimize DG contribution to ground faults
2. Sympathetic tripping: Directional relays, changes to
circuit breaker settings
3. Defeat of fuse saving: Larger fuses, minimize DG
contribution to ground faults

Impact of DG on Protection System Coordination

Figure 2. Typical feeder with recloser and fault

IV. CONCLUSION
In fig 2, IUS is the current from the utility source Fault current
seen by the recloser is IR The fault current at the fault location is
IF.
Without DG,
IUS = IR = IF
With the DG connected
IF = IUS + IDG and IR = IF
However,
IR IUS
The condition indicated is not seen in the typical radial distribution system. IUS without the DG does not equal IUS with the
DG. With the DG connected, the fault current seen by the recloser (IR) will be greater than without the DG connected.
This would normally not cause a problem with the recloser
size as long as the new greater IR does not exceed the recloser
maximum interrupting rating. However, it is very likely that
coordination between the recloser and any down-line fuses will
be lost. Because both the recloser and fuses operate faster at
higher fault currents, the required margins between the recloser
fast curve and the fuse minimum melt curve could be reduced
enough to lose coordination.
Depending on characteristics of the network and DG, various
other protection problems can arise. They are namely:
1. False tripping of feeders (sympathetic hipping)
2. Fuse coordinate with recloser fast-trip varies with DG
operation
3. Nuisance tripping of production units
4. Blinding of protection
5. Increased or decreased fault levels
6. Unwanted islanding
7. Prohibition of automatic reclosing

Different issues related to power quality when DR is integrated

with the existing power system has been discussed in the paper.
It can be concluded from this discussion that when interconnecting DR to the power system, these issues must be considered which could affect power quality and safety. Penetration of
DR can be successfully integrated with the power system as long
as the interconnection designs meet the basic requirements that
consider not only power quality but also system efficiency and
power reliability and safety.
V. REFERENCES
[1]
[2]
[3]
[4]

EPRI White Paper, Integrating Distributed Resources into Electric

Utility Distribution System Technology Review, December 2001
R.C. Dugan, M.F. McGranaghan, S. Santoso, H.W. Beaty, Electrical
Power System Quality, 2nd Edition, McGraw Hill.
C. Sankaran, Power Quality, CRC Press
Impact of Distributed Resources on Distribution Relay Protection,
IEEE Power Engineering Society

VI. BIOGRAPHY
Umar Naseem Khan was born in 1983 in Pakistan. He
received his [Link]. degree in Electronic Engineering from
Ghulam Ishaq Khan Institute, Pakistan, in 2005. After
getting more than two years of experience in Electric Power
Engineering, he enrolled in [Link]. degree in Control in
Electrical Power Engineering, Wroclaw University of
Technology, Poland, in 2007. Currently with his studies, he
is also attached with R&D Program in Electric Power
Control and Protection with Wroclaw University of Technology and Areva T&D, Poland.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
235

Using ultracapacitors for saving

energy in regenerative braking in
hybrid vehicles
Amar Patil a ,Grzegorz Maligab
a

MScs student of Departament of Computer Science, Wrocaw University of Technology

amar_ec006@[Link]

Phds student of Departament of Mechanical and Power Engineering, Wrocaw University of Technology
[Link]@[Link]

Abstract
An ultracapacitor bank control system for an Electric
Vehicle has been simulated. The purpose of this device is
to allow higher accelerations and decelerations of the
vehicle with minimal loss of energy, and minimal
degradation of the main battery pack. The control of the
system measures the battery voltage, the battery stateof-charge, the car speed, the instantaneous currents in
both the terminals (load and ultracapacitor), and the
actual voltage of the ultracapacitor. This last indication
allows knowing the amount of energy stored in the
ultracapacitor. When the car runs at high speeds, the
control keeps the capacitor discharged. If the car is not
running, the capacitor bank remains charged at full
voltage. Medium speeds keep the ultracapacitors at
medium voltages, to allow future accelerations or
decelerations. The battery voltage is an indication of the
car instantaneous situation. When the vehicle is
accelerating, the battery voltage goes down, which is an
indication for the control to take energy from the
ultracapacitor. In the opposite situation (regenerative
braking), the battery voltage goes up, and then the
control needs to store the kinetic energy of the vehicle
inside the ultracapacitor. The measurement of the
currents in both sides allows keeping the current levels
inside maximum ratings. The battery state-of-charge is
used to change the voltage level of the ultracapacitor at
particular values. If the battery is fully charged, the
voltage level of the capacitors is kept at lower levels than
when the battery is partially discharged.
Keywords:
Hybrid
vehicles,
Ultracapacitor,
Regenerative braking, energy modeling, Battery pack.

I. INTRODUCTION
Ultracapacitors are a new technology that allows to
store20 times more energy than conventional
electrolytic capacitors. Despite this important
advance in energy storage, they are still far from
being compared with electrochemical batteries. Even
Lead-acid batteries can store at least ten times more
energy than ultracapacitors How ever, they present a
lot better performance in specificpower than any

battery, and can be charged and discharged thousand

of times without performance deterioration. These
very good characteristics can be used in combination
with
normal electrochemical batteries, to improve the
transient performance of an electric vehicle, and to
increase the useful life of the batteries. Fast and
sudden battery discharge during acceleration, or fast
charge during regenerative braking can be avoided
with the help of ultracapacitors. Besides,
ultracapacitors allow regenerative braking even
with the batteries fully charged In this paper, an
auxiliary ultracapacitor bank, using a Buck-Boost
converter, has been simulated. The ultracapacitor has
a capacity of 7 Farads, a nominal voltage of 300 Vdc,
and a maximum voltage of 360Vdc. It comprises 144
units in series, each one with 1,000 Farads, and 2.5
volts dc nominal (2.7 volts maximum). The maximum
current is 400 amps, and the weight of the capacitor
bank is 45 kg. The total weight of the equipment is
estimated in 70 kg.
II. THE SYSTEM PROPOSED
The Figure 1 shows a diagram of the ultracapacitor
system proposed. The power circuit has two main
components: the converter, and the ultracapacitor
bank. The equipment is connected in parallel to the
main battery, which has 26 batteries in series (312
Vdc nominal). The capacitor voltage is allowed to
discharge until one third of its maximum voltage
(around 120 Vdc), allowing to store an amount of
112 Wh of useful energy. This apparently poor
amount of energy allows having more than 40 kW of
power during 10 seconds, which is more than enough
time for a good acceleration (or deceleration) without
detriment in the battery life. The nominal power of the
traction motor is 32 kW, and the peak power is 53 kW.
During acceleration, transfer energy from the capacitor
to the main battery. During regenerative braking, move

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
236

energy in the opposite direction. Because of the

topology of converter, the ultracapacitor never reaches
voltages higher than the battery pack (self-protection).
To do all the duties during acceleration and deceleration

correctly, a good control strategy is required. The

control strategy strongly depends on the size of the
ultracapacitor. With a large capacity, the vehicle can
run taking an almost constant battery current (the
average current). Under these conditions, the
capacitor gives all the positive and negative variations
around this average current and its voltage can
indicate when is required to increase or decrease the
average current given by the battery pack. However,
this solution is costly because ultracapacitors are
quite expensive right now. This reason forces to
install an ultracapacitor as small as possible, but large
enough to avoid battery voltages too low or too high,
and battery currents (negative or positive) too high.
Under these economical reasons, many variables need
to be measured, each one with a different priority.

III CONTROL STRATEGY

Considering the high cost of the ultracapacitors, the
total capacity in Farads has to be minimized. Then, a
more complicated control strategy is required,
because the energy stored is in this case limited.
Every variable, such as instantaneous battery voltage,
battery state of charge, instantaneous battery current,
ultracapacitor initial conditions, capacitor current, and
so, need to be sensed. The speed of the vehicle also
needs to be taken in account because when the
vehicle is going to start, all the capacitor energy will
be required. By contrast, when the vehicle run at
high speeds (more than 80 km/h), the ultracapacitors
need to be empty, to be able to receive the energy
coming from a sudden emergency stop. At medium
speeds, the ultracapacitor should have in-between
charge inside. The state of charge has to be
considered because full charged batteries do not
accept any current, and hence, under this condition,
the ultracapacitor has to be discharged (that means no

more than 15-20 % of its full capacity). Only if the

car goes to a total detention, some amount of energy
could be required. By contrast, if the battery state of
charge is poor, the ultracapacitor should keep an
amount of energy higher than under normal
conditions. The state of charge is estimated by time
integration of the battery current (positive or
negative). The system also recognizes a fully charged
battery when its voltage goes up rapidly under a
regenerative braking condition. As the energy stored
in the ultracapacitor is proportional to V 2, this
voltage gives a good indication of its remained
charge. The capacitor voltage is being controlled,
through the interaction of the other aforementioned
variables, such as the vehicle speed, and the state of
charge of the battery. The measurement of the
instantaneous battery voltage, and the sign of the load
current (positive or negative), If the battery voltage
goes up rapidly, the controller activates, and a given
amount of energy is transferred to the ultracapacitor.
This situation happens when the vehicle is running
and the brake is activated. Under this condition, the
previous state of the overall system should have kept
the ultracapacitor voltage at low levels. By contrast,
in acceleration the control operation will be activated
when the battery voltage goes down (acceleration),
and when the vehicle is going to move, or
accelerating from low speeds (positive current from
the batteries). Under these conditions, and if the
battery is not fully charged, the ultracapacitors should
be at high levels of stored energy. All the operation
described above, has to be controlled by a
microprocessor, which discriminates and takes the
appropriate decisions for each particular situation.
The best way to give each variable a right
significance, is by using a combined control. This
combined control has two levels: a Primary Control,
and a Secondary Control. The Primary Control
establishes the current reference(IREF) to be given to
the ultracapacitor for each operation condition, and
the Secondary Control generates the PWM signals for
Converter. The Figure 2 shows the Primary Control.
CAP

The first duty of the Primary Control is to keep an

adequate level of energy into the ultracapacitor. This
level of energy, or charge (REF. CHARGE in Figure
2), is calculated through the EV speed (CAR
SPEED), and the battery state of charge (BATTERY
% DOD). The block named Reference Table in
Figure 2 makes this calculation, following a criterion

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
237

shown in Figure 3. The higher the speed, the lower the

charge, and the higher the battery stateof- charge, the
lower the charge too. The shape of these curves was
estimated taking in account the time the control takes to
reach the desired ultracapacitor charge.

voltage VBATT, the ultracapacitor voltage V , the load

current to accelerate the car ILOAD, the battery current
IBATT, and the compensation current ICOMP that comes
from the energy stored in the ultracapacitor. The
settings for the battery into the control were: V =300
V , and V =360 Vdc; I =-70 A , and I =70 .
CAP

min

dc

At the same time, the ultracapacitor voltage VCAP is

measured, and the actual charge is calculated. The
error signal, between the reference charge and the
actual charge is passed through a PI control, which
evaluates the amount of current reference (IREF),
necessary to maintain the ultracapacitor bank with the
desired amount of energy. All the process explained
in the previous paragraph, to maintain an adequate
charge into the ultracapacitor, is followed
accordingly. However, if the battery voltage
(VBATT) exceeds their minimum and maximum
settings, IREF is modified. A similar action is
performed when the ultracapacitor voltage VCAP is
too low or too high. The load current (ILOAD) is also
an important reference inside the Primary Control.
When this current exceeds the maximum absolute
values set on the battery pack (IBATT), the current
reference (IREF) is also modified. Then, to take in
account all these situations, some logical rules have
been implemented. These rules are programmed
inside the Limiter Block shown in Figure 2.
IV SIMULATION RESULTS
The next oscillograms show the results obtained with
the simulation. Many different situations were
simulated. To make the results more real, use the
conclusions of some experiments with the electric
vehicle were performed on the streets. These
experiments allowed evaluating the time the vehicle
needs, and the current and voltage variations during
acceleration, and regenerative braking. These
experiments showed that the vehicle can accelerate
from 40 to 60 [km/h] in a little more than 4 seconds,
and that the maximum current taken from the battery
reaches 200 amps. When the vehicle reaches steadystate at 60 [km/h], the battery current goes down at a
constant value of around 20 [A]. With these
informations, the system proposed was simulated.
The result of this simulation (from 40 to 60 [km/h]) is
shown in Figure 4. This figure displays the battery

max

min

dc

max

Adc

When the load current begins to increase, most of the

current is taken from ICOMP, which comes from
energy stored into the ultracapacitor. As the capacitor
voltage VCAP decreases (energy is going down),
more current begin to be taken from the battery, but
when the battery reaches its limit (70 amps.), the
capacitor is forced to give all the current in excess of
70 [A]. This action produces a faster decrease in the
ultracapacitor voltage. If the capacitor voltage reaches
its minimum setting (120 Vdc with a maximum of
360 Vdc), then the vehicle should not be able to
continue accelerating, because at this point, 90% of
the capacitor energy would have been used. If the
capacitor is large enough (as in this case), this
situation will not happens, and the capacitor will be
able to end its duty. Later on, the capacitor will be
able to recover energy, once the vehicle reaches
constant speed. This case is also shown in the Figure
4. When the EV reaches constant speed (60 km/h),
the compensating current ICOMP becomes negative,
charging the capacitor. As the car is not braking, this
energy is being taken from the battery, but at a
maximum value given by the limit set by the
controller (70 Adc). Once the capacitor recovers its
energy, the battery current goes down, as shown in
the final seconds of simulations in Figure 4. Finally,
when the ultracapacitor recovers the energy, which
corresponds to the value given by the Reference
Table of Figure 2, the load current takes power only
from the battery pack. A second simulation, displayed
in Figure 5, shows a deceleration from 40 [km/h] to
stop. This regenerative action takes 2.1 seconds. The
regenerative current goes from 200 Adc to zero in the
time indicated above. Both, the battery and the
ultracapacitor, through IBATT and ICOMP, receive
this current respectively. The simulation shows that
the battery receives a current smaller than the limit of
70 Adc. This happens because the maximum voltage
allowable by the battery (360 Vdc) is reached first.
This means that the battery is almost fully charged,
and then cannot receive more than 30 Adc

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
238

approximately. For this reason, most of the current is

taken by the ultracapacitor. Once the vehicle stops,
the battery continues charging the capacitor until it
reaches its final charge (the Reference Table of
Figure 2 gives the required amount of charge). In this
case the amount of energy stored in the ultracapacitor
corresponds to a voltage of around 260 Vdc.

14th Electric Vehicle Symposium, 1996.

4. Robert G Wiegers, Donald M Blackketter, Herbert L
Hess, Modeling performance of ultracapacitor arrays
in hybrid electric vehicles, Int. J. Alternative Propulsion,
Vol. 1, No. 1, 2006

It can be noted that the battery current begins to be

positive before the vehicle reaches zero speed. This is
because the capacitor needs to have more energy at
that particular speed, and then the control begins to
charge the capacitor in advance.
V. CONCLUSIONS
An ultracapacitor bank for an Electric Vehicle has
been simulated. The purpose of this device is to allow
higher accelerations and decelerations of the vehicle
with minimal loss of energy, and minimal
degradation of the main battery pack. The control of
the system measures the battery voltage, the battery
state-of-charge, the car speed, the instantaneous
currents in both the terminals (load and
ultracapacitor), and the actual voltage of the
ultracapacitor. The simulations showed that the
control system can work properly, taking in account
all the aforementioned variables.. It is interesting to
mention that, if ultracapacitor reaches in the future, a
specific energy of at least 20 Wh/kg (at this moment
some laboratory samples reach 10 Wh/kg), it will be
possible to implement EVs with ultracapacitors only.
They could give a range of 100 kms with a 500 kgs
capacitor bank, with very short charging time and
excellent life expectancy. The EV will be able to be
fully charged in few minutes. Besides, the
ultracapacitor could last all the vehicle useful life.
REFERENCES
1. [Link], Modern Electric, Hybrid Electric and Fuel
cells Vehicles, 2004
2. Maxwell. Ultracapacitors Data sheets and technical
information for 1,000 and 2,500 Farads, [Maxwell
publications]
3. L. A. Viterna, Ultra-Capacitor Energy Storage in a
Large Hybrid Electric Bus, NASA Lewis Research
Center, 21000 Brookpark Rd., Cleveland, Ohio 44135.
8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
239

Solar energy in Spain

Dariusz Szymaski
Wrocaw University of Technology
Janiszewskiego 8 str
Wrocaw, Poland
Abstract- The article shows what kind of solar power plants
is used in Spain. It describes what technologies are used to build
this objects and how do they work.

I.

INTRODUCTION

Every country spends large amount of money for renewable

energy sources. What type of green energy a country can use,
depends on what climatic conditions does it have. For example
in places with strong winds we can develop wind energy.
Another countries, especially those lying on the south
of Europe, have great conditions for using sun in their energetic
economy.
Spain is one of the most attractive countries
for the development of solar energy (fig. 1). The Spanish
government said that they will achieve a target of 12 percent
of primary energy income from renewable energy by 2010.
This includes solar power system, which will be generating
a capacity of 3000 megawatts (MW). Spain is the fourth largest
manufacturer in the world of solar power technology
and exports 80 percent of this output to Germany.
Through a ministerial ruling in March 2004, the Spanish
government removed economic barriers to the connection
of renewable energy technologies to the electricity grid.
The widely applauded Royal Decree 436/2004 equalises
conditions for large-scale solar thermal and photovoltaic plants
and guarantees feed-in tariffs.
In Spain are three interesting power station using solar
radiation to generate eletricicty. There are PS 10, Andasol
and photovoltaic power stations.

Figure 1. Yearly total of global horizontal irradiation [kWh/m2/year] in

Spain and Portugal

Figure 2. Growth of Solar power in Spain

II. PS 10
Construction of the PS10 project, an 11 MW Solar Thermal
Power Plant in Southern Spain has been completed in 2008.
This object is located in Sanlcar la Mayor, 15 km west of the
city of Seville. This type of electric power station is called
Central Receiver System (CRS). The fig. 4 shows how does
it work . This power plant is the first solution of this type in
Europe. All technologies used in this project, like glass-metal
heliostats, a pressurized water thermal storage system and a
saturated steam receiver and turbine, have been developed by
European companies.
The heliostat field is composed of 624 heliostats each of
120 m2, with a mobile curved reflective surface which
concentrates solar radiation on a receiver at the top of a 100 m
tower. In the receiver, the steam is heated up to 40 bar (250C)
by thermal energy supplied by the concentrated solar radiation
flux.

Figure 3. PS 10 power plant (bird's-eye view)

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
240

Figure 4. Main schema of working PS 10

Next, the steam is sent to the turbine, where it produces

mechanical work, which is in the next step changed into
electricity by synchronic generator. The turbogenerator output
goes to a water-cooled 0.06 bar pressurized condenser. The
condenser output is preheated by 0.8 bar and 16 bar turbine
extractions. The output of first preheater is sent to a deaerator
fed with steam from another turbine extraction. A second
preheater is fed with steam coming from the receiver. This
preheater increases the water temperature to 245C. This
solution has a task to increase the water returning from the
drum to temperature to 247C, which is flowing out from the
condenser.
Sometimes clouds curtain the sun. In that case , the plant has
a 20-MWh thermal capacity saturated water thermal storage
system (equivalent to 50 minutes of 50% load operation). The
system is made up of 4 tanks that are sequentially operated in
order of their charge status. On the figure 4 is shown, that part
of the 250C/40 bar steam produced by the receiver is going to
the thermal storage system. When energy is needed to cover a
transient period, the energy is recovered from the saturated
water at 20 bar to run the turbine at 50% load.
This is a big structure, so to reduce the visual impact of this
building the body of the tower is rather thin when we look at
from the side. The tower has 115m total height, 18m wide and
is just 8m width. The front has to have that length because it
has to allocate the 14 m wide receiver. A large space has been
left open on the bottom of the tower to give the impression of a
lightweight structure. For the visitors there is a special platform,
which is 30 m height. From that places there is a good view of
the heliostat field lying north of the tower, and the Sevilla
photovoltaic plant (PV) south of the PS10 power plant.

Figure 5. Main schema of working Andasol

power plants, including nuclear power plant. However, in

Andasol power station steam is not produced by burning fossil
fuels but by usage of solar energy. The solar radiation is
captured and concentrated by long rows of parabolic mirrors.
The radiation from the sun is reflected from the mirrors and
concentrates on long pipes filled with water or another liquid,
which has big heat capacity and small viscosity. The heat
generated in this way is enough to produce the steam required.
Solar Millennium developed parabolic trough power plants
in Europe. The Andasol 1 plant works since autumn 2008,
Andasol 2 and 3 are currently under construction in southern
Spain. The amount of electricity this power station produces is
about 180 GWh per year. The collector surface area can be
compared to 70 football fields. This is over 510,000 m2.
When the Andasol 2 and 3 will be finished they will supply
over 200,000 people with solar electricity. The construction
will probably last about two more years. They will also
contribute to Spain's supply reliability and in particular, cover
the demand peaks in the Spanish electricity grid during the
summer months. Each power plant has an electricity output of
50 megawatts and operates with a thermal storage. The task of
this storage is the same like in the PS 10. There are two tanks
which have 36m in diameter and 14m in height. Each of them
can be filled with 28,500 tons of the storage medium. A full
thermal reservoir can continue to run the turbines for about 7.5
hours at full-load, even if it rains or long after the sun has set.

III. ANDASOL SOLAR POWER STATION

This kind of solar power plant was for first time used in
commercial operation in California in 1985. Parabolic trough
power plants have already generated over twelve billion
kilowatt hours of solar electricity, which equates to providing
12 million people with electricity for one year. In parabolic
trough power plants electricity is generated using a steam
turbine which is connected with the generator. This is the same
solution like the one which is used in conventionally fuelled
Figure 6. Andasol solar power station (bird's-eye view)

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
241

IV. PHOTOVOLTAIC POWER STATIONS

V. CONCLUSION

Photovoltaics is a method for transform a solar radiation into

electrical power. Photons from sunlight knock electrons into a
higher state of energy, creating electricity. It can be achieved
by using solar cells packaged in photovoltaic modules, often
electrically connected in multiples, as solar photovoltaic arrays,
to convert energy from the sun into electricity. The term
photovoltaic denotes the unbiased operating mode of a
photodiode in which current through the device is entirely due
to the transduced light energy.
Solar cell is a device that converts sunlight directly into
electricity, which can be used to power equipment or to
recharge a battery. The first practical application of
photovoltaics was to power orbiting satellites and other
spacecrafts, but today they are used usually as photovoltaic
modules, used for grid connected power generation. There is a
smaller market for off grid power for remote dwellings,
roadside emergency telephones, remote sensing and cathodes
protection of pipelines.
Cells are sensitive for the environment so they are packaged
usually behind a glass sheet. To achieve more power from the
photovoltaic modules we can connect them together to form
solar panels. Cells can be connected in series creating an
additive voltage. Connecting cells in parallel will yield a higher
current Although the price of modules is still too high to
compete with grid electricity in most places, significant
financial incentives in Japan and then Germany triggered a
huge growth in demand, followed quickly by production.
In last years many solar photovoltaic power stations have
been built in Spain. As of January 2009, the largest
photovoltaic power plants in Spain are the Parque Fotovoltaico
Olmedilla de Alarcon (60 MW), Planta Solar Arnedo (30 MW),
Parque Solar Merida/Don Alvaro (30 MW), Planta solar Fuente
lamo (26 MW), Planta fotovoltaica de Lucainena de las
Torres (23.2 MW), Parque Fotovoltaico Abertura Solar (23.1
MW), Parque Solar Hoya de Los Vincentes (23 MW), Huerta
Solar Almaraz (22.1 MW), Solarpark Calveron (21 MW), and
the Planta Solar La Magascona (20 MW).

The Spanish government continues to promote the

investment and expansion of both photovoltaic and solar
thermal power in the country, with a goal of 400 MW installed
power for PV and 500 MW for solar by 2010. This is still only
a small piece of the countrys total power use and total
renewable energy production.
The government, however, is committed to advance this
sector. The new requirements of 2006 requires increased
energy efficiency and an obligation to meet a significant part of
the hot water demand with passive solar heating. The Plan of
Renewable Energies sets lofty goals of 5 million square feet of
solar collectors by 2010. They new Royal Decree approved in
May 2007 improves the feed-in tariffs for both solar thermal
and photovoltaic facilities. Spanish companies and research
institutions plan to continue to be at the forefront of the
growing global field.
The president of the Spanish Photovoltaic Industry
Association, Javier Anta said that the solar industry will be a
major part of the governments goal of 20 percent renewable
energy by 2020. This is still a small percentage of renewable
power. However, its grown more than 100 percent a year in
the past few years. Anta wants to continue developing new
technology which improve using solar energy.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]

[Link] 29.02.2009
[Link] 29.02.2009
[Link] 29.02.2009
[Link] 29.02.2009
[Link]
Die_Andasol_Kraftwerke_entstehen_,lang2,109,[Link], 29.02.2009
[Link]
martin_andasol_pictures_storage.pdf, 29.02.2009

Figure 7. Solar Panels Spain

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
242

SINGLE PHASE STATCOM ITS CONTROL

ALGORTHIM
Linju Jose

Department of Electrical &Electronics Engineering

Rajiv Gandhi Institute of technology, Kottayam
Kerala, India
E-mail:linju_j @[Link]

Abstract- In this paper, a new type of single

phase static compensator (STATCOM) for low
rating used in customer side is proposed. This
new STATCOM is constructed by cascading a
full-bridge (H Bridge) voltage-source inverter
(VSIs) to the point of common coupling (PCC.)
A so-called sinusoidal pulse width modulation
(SPWM) unipolar voltage switching scheme is
applied to control the switching devices of each
VSI. A new control strategy is adopted for
compensating the harmonics and reactive
current required by the load. The proposed
STATCOM has the advantage of a fewer number
of VSIs, the VSIs being identical and extremely
fast in response to reactive power change and the
control strategy adopted shows a good response.
Index TermsStatic compensator, voltage-source
inverter, PCC, reactive power, harmonic current.
1. Introduction
The proliferation of power electronics
systems in a wide range of equipments, from home
VCRs and digital clocks to automated industrial
assembly lines and hospital diagnostics systems has
increased the vulnerability of such equipment to
power quality problems. These problems include a
variety of electrical disturbances, which may
originate in several ways and have different effects
on various kinds of sensitive loads. As a result of
this vulnerability, increasing numbers of industrial
and commercial facilities are trying to protect
themselves by investing in more sophisticate
equipment to improve power quality. Moreover, the
proliferation of nonlinear loads with large rated
power has increased the contamination level in
voltages and currents waveforms, forcing to
improve the compensation characteristics required
to satisfy more stringent harmonics standards.
Now a days the requirement for power quality
becomes more and more important to keep safety of
the electrical devices and consumer satisfaction. The
growth of the non-linear loads like the devices with
switching power supplies have increased current
harmonics, EMI problems, unnecessary reactive
power and power losses. Fast switching devices like
CMOS or IGBT transistors provide implementation
of full bridge inverters to serve as a real time
parallel compensators by bidirectional energy flow
to control and compensate reactive power and
current harmonics. Static Var Compensator can be

utilized to regulate voltage, control power factor,

and stabilize power flow [1].
Most of Var compensators employ a
combination of fixed or switched capacitance and
thyristor controlled reactance. Several Var
compensator based on a voltage sourced inverter,
known as STATCOM, have been proposed and
demonstrated [2-4].
In this paper the main goal is proposing a low cost
single phase commercial STATCOM control
topology to act as an active filter for single phase
nonlinear load by using controller using PIC
microcontroller. The control technique is useful to
modify harmonics current required by the load and
compensate reactive power due to relatively
asymmetrical unbalanced non-linear loads. The
required datas are achieved from dc bus voltage,
grid voltage, load current and inverter current. Main
tuning parameters are obtained from conventional PI
controller.
2. CIRCUIT CONFIGURATION
2.1) Statcom a simplified picture
STATCOM is a Voltage Source Inverter (VSI)
connected in shunt to the system at the point of
common coupling (PCC) as shown in Fig.1. The
load current can be thought of as having three
components active, reactive and harmonic. The
idea is to control the voltage source inverter in such
a way as to make it deliver the reactive and
harmonic currents demanded by the load so that the
grid has to supply only the active current. This
means that the grid current will be purely sinusoidal
and will be in-phase with the grid voltage on
connecting the STATCOM. Ideally, it is possible to
supply the reactive and harmonic current
requirement of the load with a VSI having a DC bus
capacitor and no external power source as the net
power supplied by the STATCOM during any given
fundamental period is zero. However, a small
amount of real power will be absorbed from the grid
practically to compensate for the energy losses in
the system.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
243

3. CONTROL STRATEGY FOR THE SYSTEM

The control strategy has mainly three sections, as
follows
1. Extraction of harmonic or reactive current
from load current
2. Generation of Unit Vectors from Grid
Voltage
3. Reference voltage for PWM generation
3.1)Control block of STATCOM
The main control block diagram is shown in the
figure [Link] required datas are sensed d from dc
bus voltage Vdc, grid voltage Vg, load current
ic(t)and inverter current. iL(t), and corresponding
unit vectors is generated which is in phase and 90o
phase shift sine and cosine waves respectively. With
respect to these voltage vectors the pulse is
generated.
Vg ic(t) iL(t) Vdc

Buffers

LPF

iL(t)

ic(t)

Unit Vector
Generation

sint

LPF
Pulse
generation

+
C/Tdc

L/Ti

Vdc*

+
-

Vc 2Vl
----------(1.1)
Term Vc is capacitor voltage and Vl is effective grid
voltage. For this the converter acts like a Boost
converter to transfer the stored energy to the load.

Fig.2. Proposed circuit configuration

cost

An inductor
Fig.1 Voltage source
is used
inverter
toas link
a grid connected
the compensator
STATCOM
voltage
to the power system. In these types of applications
PWM-VSI operates as current controlled VSI. In
Statcom the aim is to make VSI continuously track
and deliver the reactive and harmonic currents
demanded by load. This calls the use for current
controller. The VSI act as a bidirectional converter
operates as an inverter to supply the compensating
voltage to load and as a converter when charging the
capacitor. The topology adopted for the bidirectional
converter is full bridge topology [6] and a capacitor
of medium rating.
2.2) Circuit explanation
The Statcom should essentially consist of a single
phase inverter, Dc side capacitor meant for dc
voltage for inverters, filter components to filter out
high frequency components of inverter output
voltage, link inductor that links the inverter output
to the ac supply side and interface magnets(if
required)and related control blocks.
Figure2 shows the circuit diagram consisting of 4
self commutated semiconductor switches (IGBT,
MOSFET) with anti-parallel diodes which, act as a
full bridge converter. In this converter configuration,
IGBT constitute the switching devices. With a dc
voltage source (i.e. charged capacitor), the converter
can produce balanced voltage waveforms of given
frequency.
The reference signals are generated by sensing the
grid voltage, dc voltage, converter current and load
current. The sensed values are given as the input to
the controller so as to generate the reference values
for compensation.
The inverter generates a three phase voltage which
is synchronized with the ac supply, from the dc side
capacitor and links this voltage to ac source. The
current drawn by the inverter from ac mains is
controlled to be reactive with little active component
which is needed to supply the loses. Main condition
for operation of the circuit is the amount of charged
voltage on the dc link capacitors, so as to have
bidirectional current flow of current, the following
condition should be satisfied

Vg(t)

Fig.3. Control block of the Statcom controller

.
3.2) Unit Vector Generation
There is no backup or battery for the system. Taking
voltage from the grid and using it for compensation
but grid voltage cannot be pure sinusoid, it consist
of harmonics and therefore corresponding to grid
voltage the current cannot be generated. One method

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
244

for PWM generation is creating truth table for grid

voltage and then triangular comparison can be done,
but the change in grid voltage affect the truth table.
More the values in truth table more accurate will be
the signals but then memory required increases
therefore a strategy known as Unit vector generation
is adopted.

Fig 5 Block Diagram for Harmonic current extraction

Fig.4 Block diagram for unit vector generation

In this method the grid voltage is sensed and

corresponding voltage is generated whose
magnitude is always unity. Fig.4 shows the block
diagram for unit vector generation in which the peak
value of the grid voltage is sensed by sensing the
zero crossing of the cosine wave (i.e. 90o phase lag
with the grid voltage.). Each low pass filter has 45o
phase difference which create cosine and in phase
component.
3.3) Harmonic current generation:
Let us assume that unit vectors are known and the
direction of current from inverter to grid is taken as
positive.
Let the grid voltage be
(3.1)
Vgrid Vg * sin st ..
Generalised Load current be
Iload Io IpSinst IqCosst Iharmonic (3.2)
Iharmonic Iload ( Io IpSinst IqCosst ) (3.3)
Where Io=dc component, Ip=peak of active current,
Iq=peak of reactive current.
Figure 6 shows the diagram for generating the
harmonic current thats required by load for
compensation. The dc component can be neglected
as it the low value and then the harmonic current is
obtained by
Iharmonic Iload ( IpSinst IqCosst ) (3.4)
The active part is obtained by multiplying by in
phase component and the reactive part by phase lag
component.

Sint and Cost are generated unit vectors, icfa(t) and

icfr(t) are active and reactive components
respectively.
4. CONTROL ALGORITHM FOR SINGLE
PHASE STATCOM
This section describes about the control of the
proposed circuit. The flow diagram given explains
the control of Statcom for compensating the
harmonic current. The controller used is PIC
microcontroller aiming for low cost system. The
controller used here is PIC18FXXXX series.
The main concept is
1. First sense the DC bus voltage and make sure its
greater than grid voltage
2. Check if grid voltage within range, if its within
range then starts charging the capacitor and enter the
Statcom Controller loop.
At the time of start, the condition of the controller is
unknown there may be chance of starting the system
with the previous conditions therefore the controller
has to be reset first then the initialisation process
takes place. After that the controller has to be started
for that first the timers are started and then unit
vector generation takes place according to the fig.4,
in user interface the protection algorithm is given
such that over current, overvoltage, initial start-up
conditions are all satisfied according to the
conditions and then if flag is set then Statcom
controller is called after the conditions the ADC
process takes place and the parameters are displayed
through DAC and the process continues.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
245

control of transmission system,IEEE Power

Engineering Society, Summer Meeting 1994.
pp.479-6.1-8.
[5] [Link], [Link], [Link], [Link],
[Link], and [Link], New static VAR control
using force-commutated inverters, IEEE Trans.
Power [Link]., [Link].4216-4224,
Sept.1981.
[6] [Link], S.K. Leung, [Link] Hui, Henry ShuHung Chung, Circuit-level comparison of
STATCOM Technologies,IEEE Trans. Power
Electronics, [Link].4 pp.1084-1092,July, 2003.
[7]Simone Buso, Luigi Malesani, Paolo Mattavelli,
Comparison of current control techniques for active
IEEE
Trans.
filter
applications,
[Link],vol.45.,No.5,pp. 722-729 Oct.1998
[8]Bhim Singh, Kamal Al-Haddad, Ambrish
Chandra, A review of active filters for power
quality improvement,IEEE Trans. [Link].
,vol.46,No.5,pp.960-971Oct.1999
Fig.6. Flow diagram of proposed circuit

The PWM generation takes place according to the

sensed value. The switching frequency given is 10
kHz and therefore controller has to complete one
full cycle within 100sec.
5. CONCLUSION
The basic principle and controller design of a
STATCOM was studied in detail in the present
work. The methods so far used were depending
upon the synchronous reference frame, hysteresis
band control and many more. In this paper a new
control strategy was developed and the harmonic
current compensation was made .On focusing the
low cost factor the controller used is PIC
microcontroller. The control strategy shows good
results.

7. References:
[1]. Don A.G. Peedder , [Link], A parallelconnected active filter for reduction of supply
current distortion, IEE Trans. [Link]. , Vol.47,
No.5, pp [Link] 2000
[2][Link],
[Link],
[Link],
[Link], [Link], Advanced Static VAR
generator employing GTO thyristors, IEE [Link]
Power Delivery, vol.3,no.4,pp 1622-1627, Oct.1998.
[3]. [Link], [Link], [Link] and [Link],
Development of a large static var generator using
self-commutated inverter for improving power
system stability, IEEE [Link] Systems ,
vol.8,no.1,pp 371-377 Feb.1993
[4].[Link], [Link], E. Stacey , [Link],
[Link], [Link], and [Link],Development
of a +/-100MVAR Static condensor for voltage
8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
246

Transients phenomena in capacitive voltage

transformers distance protection point of view.
Rafal Kurnatowski

Abstract This paper presents problems of distance

protection due to transients in capacitive voltage
transformers (CVTs). CVT is a basic equipment which
sends information to relay about voltages in protected
system. During transmission lines faults severe transients
can occur. As a result we can obtain mal-operation of
distance protection.
I.

INTRODUCTION

Operation of distance relays with capacitive voltage

transformers meets problems. They are directly connected to
transients states of capacitive voltage transformers which
appear during transmission lines faults. In this case maloperation can occur. Relay may not recognize fault place.
Moreover its difficult for distance relay to distinguish quickly
between fault within protection zone and fault at the reach
point. Another problem is a low speed and worse accuracy of
distance protection. Therefore these problems have to be taken
into consideration.
II. CVT TRANSIENTS
A. Equivalent circuit of CVT
Main components of a generic CVT are: tuning reactor,
step-down transformer , ferroresonance suppression circuit and
a capacitive voltage divider. It consists also many additional
devices which are not important in point of view transients
phenomena. There are two important diagrams below. Figure
2 is a simplification of equivalent circuit of CVT from
figure 1.

Figure 2. Simplified model of a CVT from Figure 1

Diagram from figure 2 consists following parameters:

C sum of the stack capacitances
L,R equivalent inductance and resistance of the tuning
reactor and the step down transformer
R0 burden resistance
Rf, Cf, Lf, - parameters of anti-resonance circuit
B. Nature of transients
During line faults primary voltage of CVT collapses.
Energy stored in stack capacitances and tuning reactor need to
be dissipated. This produce a severe transients for distance
relays. These transients can have different nature. It can be
forecasted by analysis of CVT transfer function (1), where
factors A and B in nominator and denominator are functions of
parameters marked on simplified CVT diagram in figure 2.

(1)
Eigenvalues of (1) determine nature of transient. Analysis of
CVT transient nature [1], shows that induced transient can be:
Combination of four aperiodically decaying dc
components
Combination of two oscillator decaying components
Combination of one oscillatory decaying component
and two aperiodically decaying dc components
(general case)

Figure1. Equivalent circuit diagram of a CVT

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
247

C. High SIR problem

Drastic problems can be created by CVT transients
during high System Impedance Ratios. It is proved (ref. [1])
that for higher SIR the fault voltage at the reach point drops to
very small values. Also another problem appears. It is
extremely unfavorable signal to noise ratio. Its illustrated at
figure 3.

B. Ferroresonance suppression circuit

A ferroresonance suppression circuit is designed to
prevent subsynchronous oscillations. They appear when there
is saturation of the core of a step-down transformer during
overvoltage conditions. This circuit has significant impact on
the characteristic of the CVT transients because it creates an
extra path (apart from the burden) for dissipating energy.
C. CVT load
Better CVT performance is obtained when it is fully
loaded. Contemporary digital relays introduce a very low
burden tor CVT (100-400VA). As a solution we use often
artificial load.

Figure 3. Signal-to-noise ratio for the CVT transients and high System
Impedance Ratio

D. Point on wave when a fault occurs

The most severe transients appear at the zero crossing
of the primary voltage. In this case transients can reach 40% of
the nominal voltage magnitude.
IV. DISTANCE PROTECTION PROBLEM

As illustrated in Figure 3, the magnitude of the noise

components may be 10 times larger than the magnitude of the
60Hz operating signal. Noise dominates for 1.5 to 2 cycles.
Noise component frequency can be close to 60Hz in
some part of time axis. Moreover noise vector can be in
opposition to operating signal. In this case both signals cancel
mutually to certain extend. Thus the phasor estimator tends to
underestimate the magnitude.
III. CONTROL OF TRANSIENTS

Transients are basically controlled by the following factors

(reference [1]):
Sum of stack capacitances
Shape and parameters of the ferroresonance
suppression circuits
CVT burden
Point on wave when a fault occurs

A. Sum of stack capacitances

Magnitude of the transients is lower when the sum of
stack capacitances is higher. It is true that behavior of a
distance relay doesnt depend only on magnitude of fault
signal. Price of stack capacitance is quite high. Therefore
transients magnitude takes secondary importance for transient
overreach and speed of operation.

Distance relays measure impedance to determine

location of a fault. Due to transients in CVTs , voltage signal
can be underestimated. Moreover current can be overestimated.
These problems often lead to error in impedance measurement.
It results in mal-operation of distance protection.
Nowadays there are known methods of control a
current overestimation. As a solution we use mimic filter or
band-pass differentiating filter. Unfortunately problems
connected to voltage underestimation due to CVT transients
are still significant. Series of simulation show that voltage
estimation error can be equal even 90%. If we assume that
current was estimated correctly, still we have 90% of
impedance measure error.
There are a lot of methods to improve distance
protection during CVT transients. One idea is pre-filtering
voltage with non-symmetrical FIR filter. Another ideas are e.g.
to use artificial intelligence techniques or testing new
algorithms.

REFERENCES
[1] B. Kasztenny, D. Sharples, V. Asaro, M. Pozzuoli, Distance relays and
capacitive voltage transformers balancing speed and transient
overreach, 55th Annual Georgia Tech Protective Relaying Conference
Atlanta, May 2001.
[2] J. Izykowski, B. Kasztenny, E. Rosolowski, M.M. Saha, B. Hillstrom
Dynamic compensation of capacitive voltage transformers,
IEEE
Transactions on Power Delivery, Vol. 13, No. 1, January 1998

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
248

Power quality in automatic restoration systems using

stand alone generators
Piotr Danielski, Marcin Dbowski

Wroclaw University of Technology

Wrocaw, Poland

Lately automatic restoration subject become very important.

Reasons for this to happen are economic, political and
technical. For economic we could find: to big power demand
factor for production of energy ratio, negligence in
maintenance cost predictions and demanded high interest rate
on each level of energy market. For political reasons we can try
to find more restrictive energy quality directives, too big
pressure on low energy price (depending on the market energy
price is national or free). And at the end we can describe
technical reasons which in many cases are partly technical
because of their origin in one of the above groups. For instance
low maintenance expenditure effects with many failures in the
network, though insulators and lines have defined lifetime
often this period, due to other important works, is being
exceeded. Another technical problem comes with load factor in
the grid. Many faults are due to reactive power unbalance,
involving generator under-excitation, sustained overvoltage,
and switched capacitors/reactors. Other which are partly
connected with those were problems due to load and generation
unbalance, including responses to sudden increases in load and
under-frequency load shedding including lack of black-start
capability, problems with switching operation, line overloads
and control center coordination. [2]
To prevent consequences of failures many systems are
equipped in fast disconnectors and automatic restoration
system with auto reclosing procedures (for over-head lines) or
which is also common (mostly for private usage) with standby
source able to work instead of mains supply. Typical AR
system is presented on fig. 1.
[1] We can distinguish two groups of AR systems. One will
be created by systems using evident stand-by source, which are
often working with some flat period (with break after failure).
These systems are using some stand alone generators (e.g.
petrol [III.] or diesel generators, batteries and other energy
storage devices). The other group will be systems using hidden
stand-by source, so called reserve power source which is

There is also difference on the cycle depending on the failure

character. If AR is being activated because of CB operation on
main supply we are talking about shorter cycle, else if AR is
activated by the detection of voltage collapse (e.g. swell or sag)
we are talking about full cycle, which is bigger than cut off
time.

MAIN
SUPPLY

STAND-BY
SUPPLY

AR
VT

CB

INTRODUCTION

VT

(opened in normal
conditions)

I.

designed to fulfill requirements of full load even if one of the

supplying lines is cut off. In this system there are no breaks in
supply and during normal stable work lines are only loaded
with 60-70% of theoretical capacity, but during failures there is
possibility to overload system even with more than 30 % over
the maximum nominative power capacity.
What is important when talking about Automatic Restoration
are operation times (also as a factor of division). Main times of
interest are:
- operation time, the time between the detection of failure
and restoration of voltage (full or partial depending on
characteristic of the system),
- cut-off time, the time when there is no power supply in
secured circuits.

CB

Abstract The aim of this paper is to give an introduction to

automatic restoration of voltage techniques and technology. There
is also basic generator power quality test presented, with
specification of typical small power stand-by usage. All the tests
were taken with Fluke 345 PQ meter. Introduction concerns also
basic European Directives for power quality measurements and
requirements.

Figure 1. Typical evident stand-by AR system [1]

Of course both mentioned times must be coordinated with

other system operation times like protection operation times of

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
249

both plant and distribution/transmission networks as well.

During automatic restoration also number of electric motors
needed to be taken into account, when time of operation is
being chosen.
Often after voltage restoration we can observe in the system
with connected electric drives voltage drop due to associated
inrush current, then torque drop because of voltage drop,
motors stall and as a consequence either tripping of motors
protection, bus bar CB or another operation of AR. To
eliminate these faults two restrictions were introduced. First is
the level of AR power, which should be sufficient, second is
that motors are allowed to restart only with respect to their
inrush current ( cant make voltage drop 0,7 UAR ). [2]

II. DIRECTIVES
Most important documents that influence automatic
restoration parameters as well as detection parameters are
standards and normalizations. For Poland these papers are,
(taken from European Council Directives)[3] PN-EN 50160
Parameters of supplying voltage in communal distribution
networks, PN IEC 60038 Normalized voltages IEC. For
other countries respectively are being used 96/92/EC
Concerning common rules for the internal market in
electricity, and the 89/336/EEC with add-ons 92/31/EEC and
93/68/EEC Electromagnetic Compatibility Directive.
Regulations concern general requirement factors as frequency,
voltage deviation, shaping of voltage and content of unwanted
harmonics related to base harmonic and overall time of failures
in circuit.
So base values are:
- For frequency grid values for synchronized connected
circuit requirement is:
o 50 Hz 1% for 95% time during usage period
o 50 Hz 4/-6% for the remaining period
For other circuits 95% period can use threshold 2%,
and for 100% period 15%
- For voltage, average value with sampling time 10 min.
should be:
o 95% of these values should be 230/400 V
10%
o 100% of these values should be in range of
230/400 V +10%/-15%.
- For harmonics normalization is shown in table 1:
TABLE 1
Admissive values of harmonics in LV/MV networks
odd harmonics
even harmonics
multip. of 3.
not multip. of 3
Harm.
Rated
order
volt.[%]
Harm.
Rated
Harm
Rated
order
volt.[%]
order
volt.[%]
5
6
3
5
2
2
7
5
9
1,5
4
1
11
3,5
15
0,5
6..24
0,5
13
3
21
0,5
17
2
19
1,5
23
1,5
25
1,5

European Directives are also using THD factor (total

harmonics distortion) which cannot exceed 8% for networks
with nominal voltage level beneath 110kV and 3% for
networks with nominal voltage level above 110kV.

III. LABORATORY TESTS

The goal of laboratory test was to examine the behavior of
simple power generation unit as a stand-by power source.
Device chosen for testing was 1-phase generator powered by
no-PB petrol engine produced by NUAIR POLSKA Sp. z o.o.
The nominal power is 0,65 kW and the maximum power is
0,78 kW. More specific data are listed in table 2.
Not like other big machines, this one is provided with
manual starter, which is useless with auto power restoration
control system, but will give overview of advantages and
disadvantages of stand-by power source commonly used for
fulfilling needs of simple small countryside household
typically most unsecured in power supply.
Before the tests started there was found a problem with the
load. The idea of the tests was to create environment which
will be similar to real conditions. On the other hand, reduction
of the influence of inductances and capacitances was necessary.
Decision was made to take 0,5 kW halogen lamp as a load,
which is almost 100% resistance and will not cause big inrush
currents during start procedure. Additional researches were
performed with 2 computer (laptop) power supply adaptors, to
verify compatibility with energoelectronic devices.
The measurements and transients where recorded using
FLUKE 345. All the data was transferred directly do the
computer using software provided by the producer. The lamp
was connected through special socket with splited input wires
to ensure reliable usage of current clamps. For the purposes of
the test 5 connections of FLUKE inputs were used 2 current
clamps, 2 voltage probes and 1 PE wire.
The measurements were dived into three parts. All began
with cold startup. Cold startup can be understood as starting the
generator from temperature below 0C. Such strict parameters
were taken to simulate hard winter conditions the result from
pre-test is the smooth behavior during start from temperature
above 15C.
In first part after starting the generator, there was up to 10
seconds of idle work and then the load was connected.
Difference between first and second part is joined with
duration of idle work in second section it was increased to 23 minutes, this time was sufficient for engine to get to the
nominal work temperature. The last part was very similar to
second, but the load was changed to two computer power
supply adaptors connected in parallel.
From all attempts listed below values and transients were
taken:
- Current RMS,
- Voltage RMS,
- Total power,
- cos,

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
250

Total harmonic distortion (THD) parameter.

TABLE 2
NPEGG780 STAND-BY POWER SOURCE DATA

Rated power

650 W

Maximum power

780 W

Nominal voltage

Load is switched off

230 V

Output

1-phase socket

Protection

Over current (5 A)

Engine rotation speed

Engine power

3000 rpm
2 HP at 3000 rpm

Capacity of fuel tank

4,2 l

Maximum continuous work length

Level of noise

Load is switched on

6h
57 db (7 m away)

Weight

17,5 kg

Dimensions

328x325x337 (mm)

Engine type

Two stroke

Figure 3. Voltage RMS transient for warm engine [B]

B. Cold startup, 2-3 minutes of idle work

A. Cold startup, 10 seconds of idle work
This attempt will show if generators powered with petrol
engines are able to withstand load demand during power
system voltage collapse. Performances of generator after cold
start were poor. Just after the startup value of RMS voltage is
going up to almost 300V this transient lasts maximum 3
seconds, but usually ends after 1 second. Then the voltage level
is stabilizing, till the moment of connecting the load. After
connecting the load, the RMS value goes up to 250V. Almost
all of the attempts finished after 15-20 seconds, because the
generators could not withstand the load demand and stopped
working. Figure 2 shows voltage RMS value transient.
Value of cos was constant 1 for non-load period and 0.98
for load period. Just before the generator stopped it started to
decrease.
Current level was stable (2,1 Amps), only during switching
period reached 2,2 Amps but this limitation was only due to
small power of cold generator, normal conditions were
presented in [B].

The pre-test trials showed that the engine needs time to

stabilize its work. Basing on this knowledge next part tests
were performed after reaching the work temperature. After this
time the voltage level of idle work was about 230V, so it was
similar to parameters decelerated by the producer. After
switching on the load, there was o big voltage decrease, down
to 186V. The drop duration was short less then 1 second.
During work voltage average value was 240V. The transient is
not stable it is oscillating around average value. Transient of
the voltage RMS value is shown on figure 3.
The current level, like while first part, was stable (2,1 Amps).
The peak during load switching operation reached 3,3 Amps,
which is significant for other connected devices.
Value of cos were constant 1 for non-load period and
0.98 for load period. Only changes were registered when the
load was connected 0,58 for switching on and 0,91 for
switching off operation.
During this part the THD parameter was measured. It is
giving overview of harmonic content in signal. The parameter
is defined as the ratio of the sum of the powers of all harmonic
components related to the power of the fundamental frequency,
as in

Load is switched on

Engine starts to slow

down

THD =

harm. powers P2 + P3 + P4 + ... + P40

=
[%]
fund. freq. power
P1

The acceptable level of this coefficient is 20% (for separated

LV networks, including up to 40th harmonic). Obtained value
from measurements is in range around 17% for non load
period and slightly above - 21% for load period. The producer
of the generator does not give the THD value for the unit.
Figure 2. Voltage RMS transient for cold engine [A]

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
251

C. Computer power supply adaptors

The aim of last part was only to check if power source is able
to work efficiently with normal household devices. The most
interesting values, from investigation point of view, were THD
and reactive power.
Unexpectedly THD value for 2 devices connected, was very
low 2.5%.
As expected reactive power was twice bigger then active
power 90 Var. This is the result of AC-DC converter.
All other measured parameters were on expected level:
- voltage 235 V,
- current 0,42 A,
- active power 43 W.

IV. SUMMARY
The tests were performed to prove abilities of stand-by
power sources. Second topic, which should be taken under
consideration, is the control of reserve power source it should
be 100% automatic. The basic requirement for generator is the
electric starter. From the second view not all the devices and
machines connected can be connected to second power source,
the best example are the engines due to inrush currents.
Designer of control systems should pay attention to the
temperature of the engine. As it was shown during laboratory
tests it is strongly connected with efficiency and work stability.
It can be noticed that during switching load procedure on cold

engine, the peak voltage is lower, it is probably connected with

dose of fuel and poor performance of cold engine larger dose
allows engine to work against provided load.
During cold start, control system could take as a control
variable the value of cos it was observed that it is
decreasing while the generator is not able to withstand the load
demand. This theory needs further investigation.
The main problem of small generation units is the power
quality, which is decreasing in proportion to load. If the
receivers are sensitive to harmonic content, then they can not
be used.
The tests showed that even small generator unit can be used
as stand-by power source. What is more, if there is no need of
continuity of supply, the control system can be replaced by
manual control the blackout (operation) time will be
extended to few minutes.

REFERENCES
[1]
[2]
[3]

prof. B. Miedziski Lecture materials for System Opaeration and

Security. Wrocaw 2008 (not published).
[Link], M. Sarma, T. Overbye Power System: Analysis and Design.
Fourth edition (International Student Edition), USA 2008 by Thomson
Corp.
Edward Siwy, Dostosowywanie przepisw polskich w zakresie jakoci
energii elektrycznej do wymogw Unii Europejskiej Instytul
Elektroenergetyki i Sterownia Ukadw, Politechnika lska, Gliwice

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
252

Ocean Power: tidal and wave energy

Dennis Grzelak, Nico Brose, Lennardt Schnemann

The growing necessity of a sustainable energy

supply requires a great expansion of renewable
energy production. While developing and realizing
different concepts of energy usage like
photovoltaic, wind, geothermal or biomass energy,
another big potential energy source has got very
low attention energy from tidal power. Tidal
power allows many ways of electricity production
and can be classified into two different main types:

Tidal stream systems make use of the

kinetic energy of moving water to
power turbines, in a similar way to
windmills that use moving air. This
method is gaining in popularity
because of the lower cost and lower
ecological impact compared to
barrages.
Barrages make use of the potential
energy in the difference in height (or
head) between high and low tides.
Barrages suffer from very high civil
infrastructure costs, a worldwide
shortage of viable sites, and
environmental issues.

from the resultant tidal range can be calculated as

follows:

i!s the density of seawater (approximately

1025 kg/m)
A is the swept area by the rotors (m)
g is the gravitational acceleration
( 9,80665 m/s)
h is the difference between ebb and flut(m)

The best known representative of this technology is

the tidal power station in the mouth of the river La
Rance near Saint-Malo at the French Atlantic coast.
It was established already in 1966 and has a tidal
range of 12 - 16 m with which a maximum power
of 240 MW is reached. The concrete embankment
has a length of 750 m and the storage reservoir
basins magnitudes add up to 22 km in length and,
on average, about 1 km in width.

Since the economical potential of hydroelectric

power plants in Europe is nearly exhausted, this
work will be focused on different concepts of tidal
barrages, tidal stream and wave generators.
Tidal power stations are a variation of the classical
hydroelectric power plants using the level
difference of the water between low tide and flood,
also called tidal range, for the production of
electricity. For this purpose, dams can be
established in estuaries or sea bays as a separation
from the high seas. Such embankments get notches
so that water will stream in when the tide is high
and drain off when the tide is out, a process under
the driving mechanism of the turbines. In both
operation phases electric current is therefore
produced. The big advantage of this technology is
that low tide and flood continuously alternate and,
hence, the tidal force displays a computable and
dependable energy source. The potential energy

figure 1: water inflow (flood)

figure 2: water outflow (ebb)

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
253

Tidal power stations can only produce electricity at

the precise moment of the occurrence of either
flood or low tide. When the flood comes, the water
streams through the notches in the embankment
into the reservoir with the lower water level,
making the turbines turn. Being coupled to the
turbines, generators are then driven just as well,
producing electric current. As soon as the
maximum water level is reached, the streaming
comes to a standstill and no more electricity is
produced before the low tide arrives. During low
tide the water from the reservoir runs off again into
the ocean, driving the generators once more. The
turbines are generally pipe turbines, equipped with
adjustable and reversible shovels. This makes it
possible that the turbines can work in both
streaming phases. Between the flood and the low
tide streamings the equipment comes to a brief
shutdown in each case. Consequently, a
periodically variable amount of power is produced,
each streaming period lasts about 12.5 hours. In
addition the power, albeit to a minor degree, is
influenced by weather conditions like wind or
precipitation.
If suitable turbines are installed and the storage
basin is big enough, a tidal power station can also
be used as a pumping accumulator power station. In
this special case the turbines still pump water into
the storage basin beyond the reached water level,
even if the flood stream has already come to the
shutdown. Then, with low tide, more water is
blown off and therefore more electricity is
produced.
The difficulty of this energy change process lies
mainly in the fact that, as a rule, an economic
application is only possible from a tidal range of 5
m on. That is why only few locations can be
considered. Possible application places for tidal
power stations are, for example, in the mouth of the
river Severn between Britain and Wales, in Alaska
near Anchorage, at the Cambridge Gulf in Western
Australia and in the Fundy Bay in Canada.
Worldwide there are approximately at least 100
possible locations for those facilities. In Germany
the potential is estimated to be very little on
account of the relatively low tidal range of about 2 3.5 m. Furthermore the tidal power stations are also
facing the danger of corrosion, which is clearly
higher than in soft water areas. Additionally
impurities caused by salt and sediments are
problematic because they require a large
complexity in the operation management. Finally it

must be mentioned that, above all, the

establishment of big dams always makes such
equipments affect or damage nature, and therefore
interferences of the flora and fauna can be the
results. Since tidal power stations also have effects
on the rhythm of low tide and flood, an adaptation
to the changed living conditions becomes necessary
for the creatures in that sphere.

Alternatively to those
tidal power stations which
used the tidal range to
produce electricity, the
first projects are being
realised whose technology
intends the winning of
electricity from sea flows.
They are sea flow power
stations, resembling the
wind power equipments
with the difference that
the rotor turns underwater
and that it is driven by the
figure 3: Tidel turbines (horizontalconstant tidal change. A axis propeller design)
suitable pilot plant was
successfully installed already in summer, 2003 near
the coast of North Devon in England under the
project name "Seaflow". In another step, a new
equipment with a power of 1.2 MW originated in
April, 2008 near Strangford Lough under the
project name "Seagen". The principle of turbines
around which the waters are freely circulating has
the big advantage that it can be employed in
numerous locations and means only low
disturbance of the environment. Optimum locations
have a water deepness of 15 - 20 m and streaming
speeds of at least 2 - 3 m/s in the streaming
maximum with the steadiest course possible.
The power that can be extracted from marine
currents (by turbine generators) is dependent on the
speed of the water flow, and the area and efficiency
of the turbine, as follows:
Power = 0.5 *

*! A * v3 * Cp

i!s the density of seawater (approximately

1025 kg/m)
A is the area swept by the rotors (m)
v is the current speed (m/s)
Cp is the efficiency factor of the turbine.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
254

The density of seawater means that marine turbines

are significantly smaller
than wind turbines of similar generation capacity,
and rotate at approximately 20 to 30 rates per
minute.
The essential components like rotor, hub, gear,
generator and tower come from the wind power
technology. But, in contrast to the power
production from wind force, water has a much
higher density than air. The result is thousandfold
as much as the power of airflow when there is a
steady water flow, which is why the flow speeds
(relatively slow in the water), as they originate
possibly from the tides, are sufficient for the
production of electricity.
The 2-blade rotor has a solid orientation along the
direction of the flow and a pitch settlement which
allows a blade adjustment of about 180 . Therefore
the operation is possible equally with low tide as
well as with flood. Besides, the pitch settlement is
used for the line limitation and retardation of the
equipment. Rotor, gear and generator are fastened
together in a framework and can be lifted out of the
water with the help of a lift device to carry out the
maintenance.

Another concept for the use of sea flows is the

Stingray Tidal Stream generator. The Stingray uses,
in contrast to tidal turbines, no air-screw but a
compensator arm for the power production from the
sea flow. To guarantee the stability at the bottom of
the sea, a star base with four beams is applied,
which are anchored in the ground. The compensator
arm is fastened to a column on the star base. The
construction weighs approx. 180 t and it is its
simple dead weight which gives it a good
anchorage at the bottom of the sea. By the
undulation of the water the fin of the compensator
arm oscillates up and down. That is how the
pressure is generated inside a hydraulic cylinder
which is used to power a generator.
In the years 2002 and 2003 the equipment was
tested with an installed rating of 150 kW as a pilot
scheme. After building a demonstration equipment
in 2005, the enterprise Engineering Business Ltd
had arrived at the conclusion that the equipment
construction would possibly not be practicable in
consideration of profitability.

All in all, sea

flow
power
stations
could
become reliable
energy suppliers
in the future
above all due to
the computability figure 4: Stingray Tidal Stream Generator
of the tides. They
work silently and hardly disturb the landscape.
Besides, a combination with offshore wind parks
could create considerable synergy potentials, for
example, by using a common grid connection. Only
in Europe more than 100 suitable locations for such
equipments would be possible at a guess.
Nevertheless, it is problematic that suitable and
unsuitable locations are often established close to
each other. When a good location is being searched
for, it is also important to consider the possible
potential for conflicts with other ocean users
sufficiently, apart from disputable environmental
conditions. Before a commercial production of
electricity can become reality, a huge number of
detailed questions is still to be solved with regard to
the equipment technology to guarantee a safe and
frictionless operation under rough conditions, like
the sea water.

Another alternative of producing electricity is

using wave energy. At this, 3 possible technologies
are available after all:

Oscillating
Water
Column: the
periodical,
vertical
movement of
the
water
compresses
and relaxes air figure 5: Wavegen
with an airborne
turbine.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
255

Waveinduced
Drop
Height:
shafts lift
the water
on a
higher
figure 6: Wavedragon
energy
level. The potential energy is taken up by a
"low head" turbine.

Hydrodyna
mic
Movement:
buoyancy
bodies
follow the
undulation.
Out of this,
figure 7: Pelamis
the most
different possibilities of direct energy
conversion arise.

equipments running which will have a total

capacity of 18.75 MW.

The
Wavegen
project is one of
the most reliable
sea power stations.
The energy of the
water is not used
directly, but the
figure 9: Wavegen
working medium
air is. The structure is partly on shore and in the
water. It contains a hollow cavity which is flooded
by the sea. With waves streaming in, the water level
changes inside the structure. An alternately high
and low air pressure inside arises from it. The
pressure is compensated with the aid of an open end
towards the environment. Close to the open end,
there is a Wells turbine which dissipates the kinetic
energy of the flowing air into electric energy. The
Wells turbine distinguishes itself by the fact that it
maintains its running direction while the air flow is
varying, so that it supplies permanent electricity.
Before a Wavegen project can be realised,
comprehensive
investigations
about
the
characteristics of the waves have to be carried out.
An implementation of the equipment construction
has been on the Scottish island of Islay since the
year 2000.
References:

figure 8: Pelamis- Wave-Generator

The Pelamis-wave power station is a 750-t-steel

construction which generates electric current by
moving the single limbs in opposite directions. This
150-m-long and 3.5-m-wide steel tube is divided
into four cylinders which are connected by power
modules. In the modules there are piston pumps
which take up the kinetic energy and deliver it to a
hydraulic generator by means of the hydraulic
liquid. The capacity of the equipment constitutes
750 kW. An anchor holds the construction in the
intended position and is flexibly applicable with it,
in contrast to other sea power stations, and has a
lower impact on the ecological balance. The first
commercial employment occurs in 2008 in
Portugal, being part of the pilot project of 2006. In
this connection it concerns three equipments. For
the future it is planned to start another 25

Voith (2006): Meeresenergien eine Innovative

Ergnzung der klassischen Wasserkraft; Berlin, 18.
Januar 2006

Boyle, G. (2004): Renewable energy, Oxford

University Press

Lbbert, D. (2005): Das Meer als Energiequelle:

Wellenkraftwerke, Osmose-Kraftwerke und weitere
Perspektiven der Energiegewinnung aus dem Meer,
Wissenschaftliche
Dienste
des
Deutschen
Bundestages, Berlin, 10.11.2005

Bard, J. (2004): Energienutzung im OffshoreBereich, Informationen zur Raumverteilung, Heft 7/8.

2004

[Link]
[Link] , 04.03.2009

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
256

OVERVOLTAGE PROTECTION DEVICES

C. Szafron
Wroclaw University of Technology
Janiszewskiego 8 Str.
50-370 Wroclaw, Poland

Abstract This article states about overvoltage

protection methods, especially lightning overvoltages.
Different types of lightning protection devices are
presented and described, according to their
construction and application methods. The pressure
was put on protection in accordance with European
standards.
I.

INTRODUCTION

Nowadays business, industry and all public

institutions depend on electronic data engineering.
Electronic data processing, measuring and control
systems, instrumentation and control as well are all
part of a modern industrial plant. All data recording
devices at the production plants are connected to
terminals and computers at offices by information
networks distributed between buildings - together
making the computer integrated manufacturing.
Often, the basis of computer integrated
manufacturing are open networks, where different
types of computers and different operating systems
communicate. This business process is rapidly
expanding and is now approaching the computer
integrated enterprise or computer integrated
business. Differently speaking all ranges of
administration are completely integrating into a
electronic data processing system. It can be clearly
seen that the future lies in the computer-integrated
factory or in computer-integrated business and
administration. This networked system, with its
growing flow of information, is severely hindered
by damage to the essential transmission systems in
the telephone and data networks, as well as at
terminals or by interference. Strong dependence on
electronic data processing can quickly lead to
catastrophe if the system will fail. An American
investigation in 1987 highlighted the seriousness of
the situation. According to it, banks will be able to
manage without electronic data processing for only
2 days, enterprises that are sales-oriented will be
able to manage for 3.3 days, manufacturers for 4.9
days and insurance companies for about 5.6 days.
Enterprises without functioning EDP, according to
an investigation by IBM Germany, will collapse
after about 4.8 days. It has to be taken into account
that in the past 15 years the growth of computer
appliances in business, industry and everyday life
has increased rapidly, so former mentioned times
have probably decreased. Computer safety experts
have pointed out that ninety percent of enterprises
would close if their computer fails for two weeks.
Disturbance of the data flow and destruction of

Fig.1. Air terminator and equipotential bonding

electronic equipment is the most frequent reason for

the failure of such electronic systems. Mostly it is
made by transient electromagnetic interferences.
This risk can be controlled by electromagnetic
compatibility measures, what specifies conditions
under which any kinds of electric equipment
disturbs or not each other. Also where
electromagnetic
phenomena,
like
lightning
discharges, will not disturb their function. In 1989
the
European
Community
has
declared
electromagnetic compatibility as a protection goal
by issuing the Council Directive of 3 May 1989 to
Harmonise Laws of the Member Nations
concerning
Electromagnetic
Compatibility.
According to it all apparatus, facilities and systems
that include electric or electronic components must
demonstrate sufficient withstand levels against
electromagnetic disturbances, to guarantee stable
and proper operation of all equipment. The
directive of the Council especially mentions the
industrial equipment, telecommunication networks
and equipment facilities. Violation of the
electromagnetic compatibility general instructions,
is deemed with a summary offence. A great extent
the protective measures that must be undertaken in
the framework of electromagnetic compatibility.
Among the threats from the electromagnetic
environment, lightning discharge is the most
dangerous and therefore it determines the protective
measures. Modern lightning protection does not
only mean protection of buildings, but especially
the protection of devices stored in it, meaning that a
lightning protection system also must be achieved,
even if it is not necessary for the building itself, but
for the equipment it contains.
II.

AIR TERMINATIONS

To protect the volume from direct lightning strikes

and to avoid uncontrolled strikes air termination are
used [Fig.1], which are fixed points for likely

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
257

lightning strikes. Air termination is comprised by

air-termination rods and air-termination wires.
The location of air terminations is usually defined
by the rolling sphere method. It bases on use of an
imaginary sphere of radius S over the surface of a
protected object. The sphere rolls up and over and
is supported by) lightning masts, shield wires,
substation fences, and other grounded metallic
objects that can provide lightning shielding. A piece
of equipment is said to be protected from a direct
stroke if it remains below the curved surface of the
sphere [Fig.2].This means that a certain radius of
rolling sphere will be assigned to every protection
level. Finally, air terminations form a system of
protection for structures on the roof, such as
ventilators and air-conditioning systems.
Partially isolated lightning protection systems are
usually installed on flat roofs. This protection can
be achieved by individual or a combination of
several air termination rods for smaller roof
structures. For larger roof structures protection by
means of air termination rods is not often possible
as the rods would be too high and thus there is
danger of their leaning. An isolated air termination
is a second solution as an alternative.
The calculated safety distance must be smaller than
distance between air terminations and structures on
the roof. Air-termination networks must form a
protective volume that is including all structures on
the roof.
III.

BUILDING AND ROOM SHIELDS

Extended metal components, like for example:

metal roofs and facades, steel reinforcements in
concrete, expanded metals in walls, lattices, metal
supporting constructions and piping, which form an
effective electromagnetic shield by their meshed
interconnection are especially important for
shielding magnetic fields. They are also important
for creation of lightning protection zones.
In principle, a steel reinforcement and the metal
window and door frames can form an
electromagnetic cage. In practice, however, it is not
possible to weld or clamp every nodal point for
large structures. For bridging expansion joints or
bonding the reinforcement of prefabricated concrete
parts, fixed earthing terminals are provided.
To such earthing terminals or projecting bonding
conductors the earth bus or earth ring bus (ring
equipotential bonding bars) are connected. Metal
facades are also used for shielding purposes.
The facade steel sheets, being interconnected, are to
be bonded to the metal construction and to the
reinforcement.
Some of the above mentioned shielding measures
can also be applied to the establishment of room
shields (lightning protection). In particular this
concerns the use of steel reinforcements (in floors,
walls and ceilings), expanded metal (in walls and
ceilings) and lattices.

Fig.2. Rolling sphere method

Smaller shields for lightning protection zones and

higher, or shields of local lightning protection zones
are usually formed by the enclosures (sheet steel
cabinets, sheet steel covered racks, sheet steel
enclosures) of telecommunication systems and
devices.
IV.

SHIELDS FOR CABLES IN BUILDINGS

Cables shall be run near the equipotential bonding

lines. These are parts of the steel construction,
reinforced walls, cable supporting structures, cable
trays or other electrically conductive parts which
are connected to the equipotential bonding system
at least at both ends. In principle, shielded cables
should be used. This applies for electronic or data
cables as well as for higher voltage levels.
Pair-twisted signal cables are to be preferred.
For related lines of one signaling circuit a twisted
pair each is to be used, so that the incoupled
transverse voltage on cable runs must be neglected.
The limitation of the incoupled series voltage
determines the protection measures.
To lessen surges by overcoupling, power and
signaling cables must be consequently separated, if
possible by using cable supporting structures which
are included in the equipotential bonding. Different
separations are needed depending on the cable
parallel running length.
Also in existing systems it may become necessary
to shield the cable routes (subsequently). For this
purpose a retrofit set consisting of shielded
sleevings provided with a closing system in the
longitudinal direction.
V.

EQUIPOTENTIAL BONDING

Lightning protection equipotential bonding of a

protected volume includes all incoming metal
installations.
In
the
case
of
extended
telecommunication systems, a duly shaped
lightning protection equipotential bonding bar
installed at ground level inside the building also
functions as an earth bus and is usually installed as
an earth ring bus inside the building. The earth ring
bus, a ring equipotential bonding bar, is a copper
bar having a minimum cross section of 50mm2 for
surface mounting at a distance of some centimeters

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
258

from the wall. At distances about 5m it should be

bonded to the foundation earth electrode.
This bonding can also be realized over the
reinforcement. An equipotential bonding bar can be
sufficient for small local systems. If the discharge
system consists of plain metal components which
constitute an effective electromagnetic shield,
the equipotential bonding bars can be directly
bonded with the shield. Equipotential bonding is,
therefore, often carried out using a bonding plate
with multiple radial or even coaxial connections of
the conduits or line shields.
Equipotential bonding is not only for the protection
of electronic systems, but it must also fulfill special
functions. A low-impedance equipotential bonding
system, which is an entity formed of interconnected
equipotential bonding lines including the metal
parts of the electric systems (such as enclosures,
racks, cable trays etc.) and the building
(e.g. reinforcement in floors, walls [Fig.1.] and
ceilings, supporting structures between floors)
is possible using a meshed, plane-covering or
space-covering formation. Such a meshed overall
building equipotential bonding system is the best
way to reduce overvoltages in telecommunication
systems and is the basis for the coordinated use of
arresters (surge protection devices, filters, etc.).
Different types of functional equipotential bonding
systems are needed for telecommunication facilities
and systems. Metal supports, cabinets, enclosures,
cable racks etc. in rooms with telecommunication
facilities and systems must be included in the
mashed functional equipotential bonding. Another
possibility for achieving functional equipotential
bonding is to create an equipotential bonding
network by means of the metal supporting
structures between floors, It is useful to install a
local ring equipotential bonding bar which then is
connected to the earth ring bus several times.
Above all, equipotential bonding should include:
metal enclosures and
racks of the
telecommunication systems
conductors of electrical systems which do not
carry operational voltages and/or currents.
In the latter case, these include: protective
conductors of the power system, earth electrode
conductors of the telecommunication system, outer
shields of the telecommunication cables and if
necessary, the chassis terminals of the electronic
devices and systems.

TABLE 1
Comparison of arresters classes according to standards
DIN VDE 0675 Part 6/A1
IEC 37A/447/CDV
Arresters class B, for lightning protection
equipotential bonding purposes according
Arrester Class I
to DIN VDE 0185 Part 1
Arresters class C, for surge protection
purposes in the permanent installation,
Arrester Class II
especially for use in surge withstand
category III
Arresters class D, for surge protection
purposes in the mobile/permanent
Arrester Class III
installation, especially for use in surge
withstand category II
VI.

ARRESTERS

According to their range of application, surge

protective devices (SPDs) for power engineering
and for information technology can be subdivided
into two kinds: lightning current arresters and surge
arresters. SPDs are internationally standardized in
IEC 61643-1:1998-02 Surge protective devices
connected to low-voltage power distribution
systems part 1: Performance requirements and
testing methods. In this standard the SPDs are
distinguished according to test classes. It is difficult
for the user to understand this classification because
it is primarily meant for the producers of the SPDs.
A rather user-convenient SPD standardization is
included in the German DIN VDE 0675 Part 6/6A1
and 6A2. As requirements and tests of the German
standard are more severe than the international
standards, the German standard is taken as a basis
for arrester classification (Table 1).
We distinguish many kinds of arresters, generally
basing on their application:
Lightning current arresters must be able to
discharge lightning currents or considerable
parts of them non-destructively. They are
dimensioned and tested in accordance with IEC
61643-1/E DIN VDE D675 Part 6 and Part 6/A1.
Surge arresters only serve limiting overvoltages
at relatively low-energy surge currents.
VII.

SUMMARY

Application of devices mentioned in this article can

significantly increase the reliability of protected
networks and safety of people and equipment inside
protected structures .
REFERENCES
[1]

[2]
[3]
[4]

P. Hasse, Overvoltage protection of low voltage systems,

2nd edition, The Institution of Electrical Engineers, London,
United Kingdom, 2000
[Link] 09.01.2009
[Link] 09.01.2009
[Link] 09.01.2009

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
259

High-voltage direct current transmission lines

Tomasz Drobik

Wroclaw University of Technology

Wyb. Wyspianskiego 27
50-370 Wroclaw, Poland
Abstract this article treat about HVDC (high-voltage
direct current) technology.
In history of energy
transmission, crucial moment was War of Currents in
XIX century . From conflict between supporters of AC
and DC was winning Nikola Tesla. In effect, the most
popular way to transmission energy are AC transmission
lines. Despite many advantages of AC transmission, in
some conditions better is DC transmission, which demand
presence of high-voltage. The serious acceleration in
development of this technology was 1954 when was built
the first such system. Since those time, HVDC has
developed very intensive, and will be very important
technology in problem of transmission energy.
I. INTRODUCTION
Power does not rely only from voltage, but is equal of
voltage times current.
P=UI

(1)

For a given power a low voltage require a higher current

and a higher voltage requires a lower current. However, since
metal conducting wires have a certain resistance, some power
is wasted, and transfer as heat. The power losses in a
conductor are proportional to the square of current and
resistance of conductor.
P=RI2

generator in many aspects. Those reasons causes that AC

technology is today common in production, transmission and
distribution of electrical energy. However alternative current
transmission has also drawback which can be compensate in
DC links. Its the main reason why DC technology is chosen
instead AC:
- inductive and capacitive elements of lines put limits to the
transmission capacity and transmission distance
- is not possible transmission between two points of different
current frequency
Therefore electrical engineers research and applied DC
technology which doesnt have such limitation.
II. HISTORY
The first transmission of direct current was in 1882,
distance was 50 km long (distance between MiesbachMunichbut) and voltage level was only 2 kV. DC
transmission was developed by Rene Thury. This scientist
created own method, which based on series-connected
generator and was used in practice by 1889 in Italy. System
based on Thurys idea transmitted 630 kV at 14 kV over
distance 120 km. The next important

(2)

Power is also proportional to voltage, so for given power

level, higher voltage let decrease a current level. Higher level
of voltage, give us also lower power loss. Power loss can be
also reduced by decreasing resistance e.g. by increasing
diameter of conductor, but its demand higher economical
costs.
High voltage transmission is used to reduce lost of power,
but it cannot be used for lightning system and supplying
motors. High voltage level has to be adjust to receivers. In
AC are using transformers which decreasing or increasing
voltage to required level. In DC does not exist such
possibility. In those technology manipulation is possible for
more complicated way. To changing a level of voltage are
used electronic devices as mercury arc valves,
semiconductors devices, thyristors, insulated-gate bipolar
transistors (IGBTs), high power capable MOSFETs (power
metaloxidesemiconductor field-effect transistors) and gate
turn-off thyristors (GTOs).
In AC voltage conversion is simple, and demand little
maintenance. Further three-phase generator is superior to DC

Figure 1. Scheme of Thurys installation from 1889.

project was line Mountiers-Lyon in France which was

working between 1906 until 1936. Mountiers power plant had
eight generators which was connected in series. Line of
Mountiers-Lyon connected hydroelectric power plant,
transmitted 8600 kW, had 200 long kilometers and voltage
between two poles was 150 kV. In sum was built fifteen
similar systems. Other systems worked at up to 100 kV DC,

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
260

and was using to 30s. Economical and technical low

efficiency caused that Thurys systems was withdrawal, but
despite those reasons it was little commercial success.
The next era, was attempts with mercury arc valve. The
first such technology was put in 1932 by General Electric,
which tested mercury-vapor valves in 12 kV DC line. System
could convert current from 40 Hz to 60 Hz frequency. This
installation worked in Mechanicwille, New York. In 1941
existed underground DC (with mercury arc valves)
connection in Berlin, but due to war project was never
completed.
Crucial moment in development of HVDC was in 1954.
This moment began era of static mercury arc valve. This
system was created by ASEA and connected Sweden with
island Gotland. Up to 1975 had used technology based on
solid-state devices. From this time to 2000 had been lasting
era of thyristor valves. Future probably will belong to
commutated converters.

In the most popular configuration - monoplar, one

terminal of rectifier is connected with the ground. Second
terminal with potential another than ground, is connected
with transmission line.
Current flows in the earth between electrodes two stations
when not metallic conductor is installed. Its a single wire
earth return type of configuration. If not exist return
conductor, configuration may gives such problems:
- corrosion because long underground objects (e.g. pipelines),
- submerged return electrodes may participate in chemistry
reaction
- unbalanced current may disturbed magnetic field and
influent on navigations ships equipment.

III. COMPOTENTS AND RECTYFING/INVERTING

SYSTEMS

Presence of return wire can eliminate such effects. Those

line has to be connected between two ends of monopolar
transmission line. Second conductor is applied in depending
on economical, technical and environmental factors. In the
future will spread bipolar systems. Modern monopolar
systems for pure overhead lines carry typically 1500 MW. If
underground or underwater cables are used the typical value
is 600 MW.[2]

HVDC using mercury arc rectifiers but the most modern

way are thyristors. Thyristor is a solid- state semiconductor,
similar to the diode, but has particular property in control of
AC cycle. The insulated-gate bipolar transistor (IGBT) is
simpler and cheaper way of control.

The next system is bipolar transmission. In this

configurations, is used two wires with the same potential as
ground, and opposite polarity. Costs of this line is higher than
monopole, but despite economical aspect bipolar system has
advantages:

Fig 2. Simplification scheme of HVDC transmission

- return losses and environmental effect are reduced

- in case of fault in one line, current can flow by return path
in monopolar mode
- in very uncomfortable surrounding, second conductor can
be track in transmission towers, it gives possible flow power
even at the completely breaking second line.

Rectifying and inverting systems usually use the same

devices. At the AC end a set of transformers, often three
physically separate single-phase transformers, isolate the
station from the AC supply, to provide a local earth, and to
ensure the correct eventual DC voltage. The output of these
transformers is then connected to a bridge rectifier formed by
a number of valves. The basic configuration uses six valves,
connecting each of the three phases to each of the two DC
rails. However, with a phase change only every sixty degrees,
considerable harmonics remain on the DC rails. An
enhancement of this configuration uses 12 valves (often
known as a twelve-pulse system). The AC is split into two
separate three phase supplies before transformation. One of
the sets of supplies is then configured to have a star (wye)
secondary, the other a delta secondary, establishing a thirty
degree phase difference between the two sets of three phases.
With twelve valves connecting each of the two sets of three
phases to the two DC rails, there is a phase change every 30
degrees, and harmonics are considerably reduced.[1]. In
elements which take share in conversion, are applied filters
which limit harmonic in DC cycle.
IV. CONFIGURATIONS

Bipolar systems may carry as much as 3000 MW at voltages

of +/-533 kV. Submarine cable installations initially
commissioned as a monopole may be upgraded with
additional cables and operated as a bipole.[1]
Back to back is configuration with both inverters and
rectifiers are in the same area. This system is used as follows
areas:
- in Japan, where occur necessity of changing current
frequency
- in lines with different phase relationship
- in lines with different frequency and phase number
System with transmission line is the most common
configuration. In this structure two inverter or rectifier
stations are connected by powerline. This system is applied in
long lines, unsynchronized grids and underwater connections.
Tripole: current modulating control. Its new idea of
transmission DC (since 2004). This structure based on two
circuit which work as bipole and third wire which operate as
parallel monopole. Parallel monopole relive current in periods
from other pole. Bipole wires are loaded for few minutes. In
tripole system can be carry higher current. Tripole system let
pick up about 80% transferred power than AC lines.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
261

V. ADVANTAGES AND DISADVANTAGES

The advantage of HVDC than AC is ability to
transmission big amount of power on long distances with
lower wastes. DC technology is better in such situations:
- undersea connections
- power transmission and stabilization between
unsynchronized AC distribution system
- connection generating plants remote from power grid
- stabilizing AC grid
- connection between countries with different current
frequency/voltage
- synchronize AC produced by renewable energy sources
Long underwater lines have a high capacitance. In AC
transmission is required process of charging and discharging,
what is causes of power losses. HVDC has minimize this
effect. In AC lines occurs also dielectric losses.
To disadvantages of HVDC we can include conversion,
switching and control. Static converters are expensive. In
short distances losses in static inverter may be even bigger
than in AC transmission. In the future static converters will
be replaced by thirystors.
In DC system controlling of multiterminal configuration is
quite hard, because required good communication between
terminals. Also circuit-breakers are difficult than AC.
VI. ECONOMICAL ASPECTS
Is not simple estimate a cost of buildings HVDC
transmission line and operations unit. Cost are very different
and depends on power of line, length connection,
environment of track wires (air or water) and so on. Usually
the biggest producers of high-voltage direct current
transmission as Areva, Siemens or ABB dont reveal
financial information about investments. Despite narrow
basement of information, we can estimate some costs of built
DC line, which is show on Fig.3

HVDC system can connected unsynchronized grids,

therefore such lines very often are natural boundaries
between countries. DC line is also meeting in places where is
require undersea transmission (e.g. wind farm) and between
two long distant points. From such reasons was built grids
e.g. in Siberia, Canada, Australia or Scandinavia. Problem of
synchronized AC because of different frequency system
occur e.g. in Japan, North America, South American
(enormous hydroelectric power plant) between Brazil and
Paraguay In Europe the most lines are between UK,
Scandinavia and continental Europe.

Fig 4. Nicolet convert station in transmission line Qubec - New England

VIII. CONCLUSIONS
Today HVDC is very important issue in transmission
energy. In near future this technology probably will be
develop very intensive. Influence on future may have
intensive spread of renewable energy source, also wind farm
which need undersea connections. Also problem of cascade
blackout, can be reduced by application of HVDC. Intensive,
very large investments in e.g in China and India shows that
high-voltage direct current will very important in the future,
especially in big, new-industries countries.
VIII. REFERENCES
[1]
[2]

[Link]
G. Czisch, A "Low Cost but Totally Renewable Electricity Supply for a
Huge Supply Area a European/Trans-European Example
[3] [Link]
[4] [Link]
[5] Roberto Rudervall J.P. Charpentier Raghuveer Sharma, High Voltage
Direct Current (HVDC)Transmission Systems Technology Review
Paper

Fig.3 Relation between length of line and costs in AC and DC line

About 50% cost of DC structure are converter transformers,

valves and infrastructure as buildings.
VII. APPLICATIONS
8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
262

Transient response of doubly fed induction generator

under voltage sag using an accurate model
Alireza abbaszadeh*, Saeed lesan*, Vahid mortezapour**
[Link]@[Link]
* Electrical Engineering Department-Iran Babol (Noshirvani) University of Technology
** School of ECE University of Tehran
Babol*, Tehran**, Iran
Abstract- In order to investigate the transient response of
doubly fed induction generator (DFIG) under grid
disturbances, an accurate model is required. In this paper
transient response of DFIG during voltage sag considering
saturation effect of leakage flux for different order models is
simulated. Then, rotor over-current due to voltage sag and
factors that affect it, is investigated.

NOMONCLATURE
Symbols
Effective phase
current
Current
Voltage
Resistance
Reactance
Flux
Angular velocity
Combined inertia of
wind turbine and
DFIG rotor
Current ratio
Saturation factor

I
i
u
R
X

K
Subscript

m
d
q
r
s
M
cb
max
sat

Mutual
d axis component
q axis component
Rotor
Stator
Mechanical
Crowbar
Maximum
Saturation
Base
Leakage

l
I.

increasing. Nowadays, the most widely used wind turbine in

wind farm is based on doubly fed induction generator
(DFIG) due to noticeable advantages: the various speed
generation, the decoupled control of active and reactive
power and the improvement of the power quality. However,
wind turbines based on the DFIG are very sensitive to grid
disturbance, especially to voltage dips [1]. in recent years,
many papers investigated it from various aspect. For
studying a DFIG, at first an accurate model is needed. In
some papers different order model was proposed [2-5]. The
computational time can be ameliorated by reducing the
order of the generator. The reduce order is obtained by
neglecting differential term in the voltage equation of the
machine [5].in these papers the effect of leakage flux
saturation is usually neglected. In other kind of papers
saturation effect was considered, but investigation doesn't
perform for different orders. X s , and X r arent constant
in these papers, but vary depend on current pass through
them [6,7]. According to mentioned above, the voltage dip
could cause over-voltage and over-current in rotor winding
and consequently damaged the rotor side converter. So
studying the transient behavior of DFIG during the three
phase voltage dip is very important.
The aim of this work is to present a comparative study
among different orders of the DFIG model accompany with
saturation effect during three phase voltage sag and at next
step the most accurate model was used to show rotor overcurrent. Finally various factors reducing the maximum value
of over-current is investigated.
II.

unsaturated induction generator model

in this section the unsaturated model for DFIG is given.

Introduction

Due to the increasing concern about CO2 emissions,

renewable energy systems and specially wind energy
generation have attracted great interest in recent years.
Large wind farm have been installed or planed around the
world and the power rating of the
wind turbine is

Figure 1. direct (d) and quadrature (q) representation of induction machine.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
263

The voltage equations for the induction generator are

given below, where all quantities except the angular
frequency, b , are in pu.
u ds Rs ids qs

1 d ds
b dt

(1)

u qs Rs iqs ds

1 d qs
b dt

(2)

u dr Rr idr (1 m ) qr

1 d dr

0 dt

(3)

1 d qr
0 dt

(4)

u qr Rr iqr (1 m ) dr

Table 1.
simplifications used in different order models
Fifth order

Third order

First order

d sq ( d ) dt 0

d sq ( d ) dt 0

d sq ( d ) dt 0

d rq ( d ) dt 0

d rq ( d ) dt 0

d rq ( d ) dt 0

d m dt 0

d m dt 0

d m dt 0

III.

saturated induction generator model

for better representation of the induction machine under

transient condition, saturation effect should also include the
variation in the stator and rotor leakage inductances due to
saturation in the leakage flux path. The effect of saturation
in leakage flux path is considered by modifying the
unsaturated stator and rotor leakage reactances with
saturation factor K sat [7]. K sat is introduced as follow [6]:

Using these equations, a model for both the squirrel cage

and wound rotor generators can be developed. the difference
between these two generators lies in the rotor: the first one
has short circuited winding, so rotor voltage is zero, and the
second one has these winding fed by a converter,
responsible of controlling the generator[5]. Flux linkage
used in previous equation, is obtained from (5) and (6):

Where is equal to the ratio between current at which the

qs X s iqs X m iqr

ds X s ids X m idr

(5)

saturation begins, I sat , to the current through the leakage

reactance. The saturation factor characteristic is shown in
fig.2.

qr X m iqs X r iqr

dr X m ids X r idr

(6)

Te ds iqs _ qs ids

(7)

Finally if TM is the mechanical torque dependent upon the

local wind speed:

d M
TM _ Te 2 H
dt

0.8
0.6
0.4
0.2

0
0

(8)

The model of an induction generator can have various

orders, such as 1, 3 or [Link] fifth-order model is considered
to be a full order model for an induction generator. The
third-order model ignores the stator dynamics and is widely
used in power system transient stability analysis. The first
order model ignores both the stator and rotor dynamics. The
only differential equation left is the swing equation. This
model is suitable for long term power system dynamic study
including the induction motor load characteristics [4]. Table
1 briefly describe above sentences [5].

(9)

saturation factor

The electromagnetic torque is calculated by

K sat

1
I I sat

1
(tan (
) 1 ) I I sat

10

15

20

25

30

current(pu)

Figure 2. saturation factor characteristic

IV.

transient response of DFIG during voltage sag

short duration under voltages are called "voltage sags" or

"voltage dips". The latter term is preferred by the IEC.
Within the IEEE, the term voltage sag is used. According to
the IEC, a supply voltage dip is a sudden reduction in the
supply voltage to a value between 90% and 1% of the
declared voltage, followed by a recovery between 10 ms and
1 minute later. For the IEEE a voltage drop is only a sag if
the during sag voltage is between 10% and 90% of the
nominal voltage [8]. In this paper a 50% voltage sag is
considered for simulation. At t = 5s, a voltage sag happens.
Then at t =7s the voltage is restored to its presage value, i.e.
1pu.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
264

25

16

5th sat
5th unsat
3rd sat
3rd unsat
1st sat
1st unsat

15
10

5th sat
5th unsat
3rd sat
3rd unsat
1st sat
1st unsat

14
12

Torque (pu)

Torque (pu)

20

10
8
6
4
2

0
0
-5

5.02

5.04

5.06

5.08

-2

5.1

7.1

7.2

7.3

Time (s)

(a)

(b)

35

7.4

20
15

-10

ids (pu)

25

10
5

5th sat
5th unsat
3rd sat
3rd unsat
1st sat
1st unsat

-20

-30

0
-5

5.02

5.04

5.06

5.08

-40

5.1

7.5

8.5

Time (s)

10
5

15
10

iqs (pu)

15

iqs (pu)

10

20

5th sat
5th unsat
3rd sat
3rd unsat
1st sat
1st unsat

20

5
0

-5

-5

-10

-10

-15
5

5.02

5.04

5.06

5.08

-20

5.1

5th sat
5th unsat
3rd sat
3rd unsat
1st sat
1st unsat
7

7.5

8.5

Time (s)

1.02

1.35

1.015

1.3

Speed (pu)

1.4

1.01
1.005

5th sat
5th unsat
3rd sat
3rd unsat
1st sat
1st unsat

1
0.995
0.99
5.02

5.04

Time (s)

(g)

9.5

10

(f)

1.025

Time (s)

(e)

Speed (pu)

9.5

(d)

25

0.985

Time (s)

(c)

-15

7.5

10

5th sat
5th unsat
3rd sat
3rd unsat
1st sat
1st unsat

30

ids (pu)

Time (s)

5.06

5.08

5th sat
5th unsat
3rd sat
3rd unsat
1st sat
1st unsat

1.25
1.2
1.15
1.1
1.05

5.1

7.5

8.5

9.5

Time (s)

(h)

Figure 3. DFIG transient response due to 50% voltage sag. (a) torque at voltage sag occurrence, (b) torque at voltage recovery,
(c) ids at voltage sag occurrence, (d) ids at voltage recovery, (e) iqs at voltage sag occurrence, (f) iqs at voltage recovery, (g)
speed at voltage sag occurrence, and (h) speed at voltage recovery

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
265

10

1.

after the fault occurred:

the effect of leakage flux saturation is to increase
the current and torque.
In spite of fifth order model there isnt any
oscillation in third and first order model.
Difference between saturated and unsaturated of
first order model is negligible, but for higher order
model this is significant.
The rotor speed decreases for saturated and
unsaturated of fifth order model, but for other
models there isnt any decrease in speed.

2.

after the fault cleared:

fifth order saturated model reach its steady state

faster than the other models.
Saturation doesnt affect the steady state value of
torque, current and speed but it affects transient
response of these variables.

V.

rotor over-current due to voltage sag

The abrupt drop of the grid voltage causes over-voltages and

over currents in the rotor windings that could even destroy
the converter if no protection element is included. In [9] the
factors that affect over-current was investigated
mathematically. In this paper the effect of these factors is
simulated considering an accurate model. For all simulation
in this section fifth-order model considering saturation
effect is used.
A. crowbar resistance
The first factor is rotor resistance. If we neglect the stator
resistance, the amplitude of over-current is proportional to
1/Rr. So the solution to protect converter is to short circuit
the rotor windings with the so-called crowbar resistance.
There are two main requirements that give an upper and a
lower limit to the resistance [7]:
1) the resistance should be high to limit the short
circuit current .
2) on the other hand, it should be low to avoid a too
high voltage in the rotor circuit.
A too high voltage can result in break down of the isolation
material of the rotor and the converter. Fig.4 shows a
simulation that uses crowbar resistance in series with rotor
resistance as a solution to reduce rotor over-current.
Simulation was done for different value of crowbar. When
the voltage sag occurs the crowbar resistance is added to
rotor resistance.

35

Rcb=0
Rcb=0.02 pu
Rcb=0.03 pu
Rcb=0.04 pu

30

Rotor current (pu)

The effect of saturation using different orders on the

transient performances of the generator associated with such
events is shown in fig.3. the machine used for simulation
was a 2 MW, 690 V wind turbine and machine parameters
are given in appendix.
The following results can be seen from above simulation:

25
20
15
10
5
0

5.02

5.04

5.06

5.08

5.1

Time (s)

Figure 4. rotor over-current due to voltage sag for different crowbar

resistance value

As shown in fig.4, the more the value of crowbar

resistance, the less the maximum of rotor current. Table 2
shows the maximum of rotor current while the crowbar
resistance changes.
Table 2
maximum of rotor current while the crowbar changes

Rcb ()

0.02

0.03

0.04

I r max ( A)

34.12

20.96

17.51

15.02

B. depth of voltage dip

The second factor is depth of voltage dip. In fact, the more
severe of voltage drop, the bigger of the rotor current
amplitude was. To quantify sag magnitude in radial systems,
the voltage divider model, shown in fig.5, can be used. The
voltage at the DFIG terminals can be found from (10).
Vsag

ZF
E
ZS ZF

(10)

Any fault impedance should be included in the feeder

impedance Z F .

figure 5. voltage divider model for a voltage sag

It's seen from (10) that the sag becomes deeper for faults
electrically closer to the DFIG (when Z F becomes smaller),
and for systems with a smaller fault level (when Z S
becomes larger). In (10) E=1 pu and Z F z l , with z the
impedance of the feeder per unit length and l the distance
between the fault and the pcc, leading to (11).

Vsag

zl
Z S zl

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
266

(11)

The sag magnitude as a function of the distance to the fault

has been calculated for a typical 11 kV overhead line,
resulting in fig.6.

3.

4.
sag magnitude (pu)

0.8

0.6

0.4

750MVA
200MVA
75MVA

0.2

Appendix

0
0

10

20

30

40

50

distance to the fault (km)

Figure 6. sag magnitude as a function of the distance to the fault, for faults
on an 11 kV, overhead line.

The fault level is used to calculate the source impedance at

the pcc. In this paper, 11 kV overhead line, and 200MVA
fault level is considered. The impedance of overhead line is
0.117 + j0.315 per km[8]. Simulations are done for
different distance between fault and pcc.
30

20

Rated power (W)

Rated voltage (V)
Rated frequency (Hz)
X m (pu)

2000
690
50
3.188

Rs (pu)
Rr (pu)
X ls (pu)
X lr (pu)

0.00562
0.00575
0.0708
0.0486

X S X m X ls
X r X m X lr

Vsag=0.847 pu,dis=10km
Vsag=0.735 pu,dis=5km
Vsag=0.357 pu,dis=1km
Vsag=0.217 pu,dis=0.5km

25

Roror current (pu)

and fifth order model reaches faster than the

reduced orders.
it's crucial to incorporate the saturation effect to
study the transient performance of DFIG, but it's
not important for steady state study.
in the case of strong voltage sags, the induced
voltage on rotor can even be higher than the stator
voltage. In this situation, there appear over-current
whose amplitude depends on the depth of the dip,
the machine parameters, and the protection
element.

REFERENCE

15
10
5
0
5

5.02

5.04

5.06

5.08

5.1

Time (s)

Figure 7. rotor over-current due to voltage sag for different distance

between fault and pcc.

Fig.7 shows that the distance between fault and terminal of

DFIG is an important factor for maximum amplitude of
rotor current. For example when it is 500m, rotor maximum
current is 25.8A, while it is 10km, rotor maximum current is
2.85A.
VI.

Conclusion

In this paper transient performance of different order model

of DFIG considering or ignoring saturation effect was
compared. Then some parameters that affect rotor overcurrent due to voltage sag was simulated. The main findings
can be summarized in the following points:
1. the stator and rotor current of DFIG calculated by
considering the saturation effect are significantly
higher than the ones that ignores it.
2. the rotor speed of saturated model reaches the
steady state value faster than unsaturated model,

[1] Jesus Lopez, Pablo Sanchis, Xavier Roboam and Luis Marroyo,
"Dynamic behavior of the doubly fed induction generator during threephase voltage dip", IEEE Transactions on energy conversion, vol.22, no.3,
September 2007.
[2] [Link], [Link], and [Link], "Comparison of 5th and
3rd order machine models for doubly fed induction generator (DFIG) wind
turbines", Electric Power Systems Research 67 (2003), pp.207-215
[3] Andres Feijoo, Jose Cidras, and Camilo Carrillo, "A third order model
for the doubly-fed induction machine", Electric Power Systems 56 (2000),
pp.121-127
[4] Zhixin Miao, and Lingling Fan, "The art of modeling and simulation of
induction generator in wind generation applications using high order
model", Simulation Modeling Practice and Theory, 2008
[5] Miguel Garcia-Gracia, [Link] Conech, Jesus Sallan, and Andres
Llombart, "Modeling wind farms for grids disturbance studies", Renewable
Energy 33(2008), pp.2109-2121
[6] Hony [Link], and Narayan [Link], "Effect of main and leakage flux
saturation on the transient performances of doubly-fed wind driven
induction generator", Electric Power Systems Research 77 (2007), pp.10191027
[7] Hony [Link], and Narayan [Link], "Leakage flux saturation effects on
the transient performance of wound rotor induction motors", Electric Power
Systems Research 78(2008) 1280-1289
[8] Math [Link], Understanding power quality problems, Chalmers
University of Technology Gothenburg, Sweden, 2000.
[9] Aoyang Han, Zhe Zhang, and Xianggen Yin, "Study of factors affected
the rotor over-current of DFIG during the three phase voltage dip", DRPT
2008, 6-9 April 2008.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
267

Reverse electrodialysis for Power Generation

Sameer Deshmukh
Wroclaw University of Technology
Wroclaw, Poland
Abstract- The use of reverse electrodialysis, to derive
energy from the difference between the chemical potentials
of concentrated and dilute salt solutions is reviewed. The
process employ the flows of brine and dilute solutions
through alternating cells bounded by cation- and anionexchange membranes in a stack placed between two
electrodes to generate a voltage by the passage of salt
through the membranes. Potential sources of brine include
the salt domes of oil and gas wells, salt water lakes, and
geothermal brines. This paper also reviews a range of
membrane properties of the commercially available
membranes.
Keywords: Power generation, Reverse electrodialysis,
membranes for RED.
1. INTRODUCTION
Renewable and sustainable energy sources are playing an important
role in the 21st century and are becoming increasingly
important due to environmental problems, such as global pollution
phenomena and global warming. Membrane technology provides an
opportunity to gain renewable and sustainable energy from salinity
gradients via, e.g. pressure retarded osmosis and reverse
electrodialysis. The latter method seems more attractive for power
generation using sea and river water. Reverse electrodialysis (RED) is
a non-polluting, sustainable method for energy generation by mixing
fresh and salt water. RED converts the free energy generated by
mixing the two aqueous solutions into electrical power and the
system can be applied wherever two solutions of different salinity are
mixed, e.g. where river water flows into the sea. The principle of
RED was first proven by Pattle, who with his pioneering work in
this field was the first one that generated power using RED. In the
70s, Weinstein and Leitz investigated the effect of solution
composition on power output and the main conclusion of this work
was that large-scale energy conversion by RED may become practical,
but only with major advances in the manufacturing of ion exchange
membranes and with careful optimization of the operating
conditions. In the early 1980s, Lacey conducted a comprehensive
investigation on RED. Conclusion of his work was that to make RED
commercially available it is necessary to minimize the internal stack
resistance of the RED cells and to maximize the net power output
from the cell. Lacey concluded that membranes for RED should have
a low electrical resistance and a high selectivity combined with a
long service life time, acceptable strength, dimensional stability
and low-cost. In the mid 1980s, Jagur-Grodzinski investigated
membrane spacer modifications and different salt solution streams
in order to generate more energy. In 2007, Turek and Bandura
studied the effect of solution velocity on cell power output and process

economy. Turek, mentioned that the main bottleneck for a successful

RED system seems to be the membrane price. Despite that remark,
the focus of most of the earlier work on RED was on stack design,
solution flow and solution composition, but not on membrane
characterization and membrane performance. Most scientists used
electrodialysis membranes to study the RED process. The ion
exchange membranes are the key elements in the RED system. The
most important membrane properties for RED are ion exchange
capacity, swelling degree, and of course the membrane resistance
and its selectivity (the ability of the membrane to distinguish
between cations and anions), because of their direct effect on the
overall RED performance. Up to now, it is still not known which
membrane parameter is the dominant factor with respect to power
generation in RED. It is essential to identify the key membrane
parameters for RED in order to further improve the power output
of RED. The data available in the existing literature do not offer
sufficient information on membrane properties relevant to RED to
enable proper mutual comparison of the different commercially
available membranes. Currently, no complete overview is available
with respect to the application of ion exchange membranes in RED
which covers the full range of membrane types.
This review presents an overview of reverse electrodialysis. It
investigates a range of membrane properties of commercially
available membranes that are important with respect to application in
RED. The objective of this study is to analyze the membrane
properties under equivalent conditions to enable a fair comparison
of the results and a proper evaluation for application in RED.
2. Theoretical background
2.1. Principle of reverse electrodialysis

In RED, a concentrated salt solution and a less concentrated

salt solution are brought into contact through an alternating series
of anion exchange membranes (AEM) and cation exchange membranes (CEM) (Fig. 1,1.1).
The membranes separate the concentrated solution from the
diluted solution and only ions can pass through the ion selective
membranes. Anion exchange membranes contain fixed positive
charges which allow anions to permeate through the AEM towards
the anode and cation exchange membranes contain fixed negative
charges which allow cations to be transported through the CEM
towards the cathode. The difference in chemical potential between
both solutions is the driving force for this process. At the electrodes
a redox couple is used to mitigate the transfer of electrons. The
chemical potential difference generates a voltage difference over
each membrane.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
268

2.2. Ion exchange membrane properties

Ion exchange membranes are membranes with fixed anionic

or cationic exchange groups that are able to transport cations or
anions. The specific properties of ion exchange membranes are all
related to the presence of these charged groups. Amount, type and
distribution of ion exchange groups determine the most important
membrane properties. Based on the type of fixed charge group, ion
exchange membranes can be classified as strong acid and strong
base, or weak acid and weak base membranes. Strong acid cation
exchange membranes contain sulfon groups as charged group. In
weak acid membranes, carboxylic acid is the fixed charged group.
Quaternary and tertiary amines are the fixed positive-charged
groups in strong and weak base anion exchange membranes,
respectively.
2.2.1. Ion exchange capacity

The ion exchange capacity (IEC) is the number of fixed charges

inside the ion exchange membrane per unit weight of dry polymer. The ion exchange capacity is a crucial parameter which affects
almost all other membrane properties. The IEC is expressed in milli
equivalent of fixed groups per gram of dry membrane (mequiv./g
membrane).
2.2.2. Fixed charge density

Ion exchange membranes contain fixed-charged groups attached

to the polymer backbone. In cation exchange membranes, the fixed
negative charges are in electrical equilibrium with the mobile
cations (counter-ions). The opposite relation exists in anion exchange
membranes. The fixed charge density, expressed in milli equivalent
of fixed groups per volume of water in the membrane (mequiv./l)
strongly depends on the IEC and the swelling degree of the
membrane: in the swollen state, the distance between the ion
exchange groups increases thus reducing the fixed charge density.
The transport of counter ions through the membrane is determined
by the fixed charge density in the membrane and the difference
between the concentrations of the electrolyte solution in contact
with that membrane. The concentration and type of the fixed
ionic charges determine the permselectivity and the electrical
resistance of the membrane.
The theoretical value of the potential over the membrane for an
aqueous monovalent electrolyte (e.g. NaCl) can be calculated using the
Nerst equation (Eq.(1)):

where Vtheo is the theoretical membrane potential for a 100%

selective membrane (V), R is the gas constant (8.314J/(mol K)), T
is the absolute temperature (K), z is the electrochemical valence, F
is the Faraday constant (96,485 C/mol), ac is the activity of the
concentrated solution (mol/l) and ad is the activity of the diluted
solution (mol/l). For fresh water (0.017 M NaCl, y = 0.878) and sea
water (0.5 M NaCl, y = 0.686), the theoretical voltage difference
per membrane is 80.3 mV. The overall potential of the system is the
sum of the potential differences over each pair of membranes.

2.2.3. Permselectivity

When an ion exchange membrane is in a contact with an electrolyte (salt solution), ions with the same charge (co-ions) as the
fixed ions are excluded and cannot pass through the membrane,
while the oppositely charged ions (counter ions) can pass freely
through the membrane. This effect is known as Donnan exclusion
[14]. The permselectivity of a membrane describes the charge selectivity of the ion exchange membrane. It reflects the ability of the
membrane to discriminate between ions of opposite charge.
2.2.4. Electrical resistance

The electrical resistance of the membrane is an important property of ion exchange membranes, because it is directly related
to the maximum power output in reverse electrodialysis and the
energy consumption in electrodialysis processes. The membrane
resistance is determined by the ion exchange capacity and the
mobility of the ions within the membrane matrix. The electrical

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
269

resistance is dependent on temperature and decreases with

increasing temperature. The specific membrane resistance is in
principle reported in 2 cm. However, more useful and most often
reported in literature is the membrane resistance in 2 cm2.
2.2.5. Heterogeneous and homogenous ion exchange membranes

Ion exchange membranes can be divided with respect to

their structure and preparation procedure into two categories:
homogeneous and heterogeneous membranes. In homogenous
ion exchange membranes the fixed charge groups are evenly
distributed over the entire membrane matrix. Homogenous
membranes can be manufactured via polymerization and polycondensation of functional monomers (e.g. fenylosulfonic acid
with formaldehyde) or via functionalization of a polymer by for
example dissolving the polymer in a suitable solvent and subsequent
functionalization by, e.g. post-sulfonation. Heterogeneous
membranes have distinct macroscopic uncharged polymer
domains of ion exchange resins in the membrane matrix. This
type of membranes can be produced by melting and pressing of
a dry ion exchange resin with a granulated polymer (e.g.
polyvinylochloride). Another method to prepare heterogeneous
membranes is dispersion of the ion exchange resin in a polymer
solution. The distinct difference between homogenous and
heterogeneous ion exchange membranes also influences the
properties of the specific membrane.

Here I is the current (A), Rload is the load resistance (2), Rstack is the
stack resistance (2) and Vo is the stack open circuit potential (V).
To generate a maximum power output (Wmax), Rload needs to be
equal to Rstack.
In that case, Eq. (4) changes into Eq. (5) which shows the relationship between the open circuit potential, the maximum power
output and the stack resistance.

Combination of Eq. (5) with Eqs. (2) and (3) finally yields Eq. (6)
which relates the maximum power output of the RED stack to the
individual contributions of each component. More specifically, it
relates the maximum power output of the system (Wmax) to the
average membrane selectivity (av) and the membrane resistance
(Raem and Rcem).

where N is the number of membrane pairs, av is the average membrane pair permselectivity, R is the gas constant (8.314J/(mol K)),
T is the absolute temperature (K), F is the Faraday constant
(96,485 C/mol), ac is the concentrated solution activity (mol/l), ad
is the diluted solution activity (mol/l), Raem is the anion exchange
membrane resistance (2 m2), Rcem is the cation exchange mem2.3. RED membrane model
brane resistance (2 m2), A is the effective membrane area (m2), dc
To predict the relative contribution of the different components is the thickness of the concentrated compartment (m), dd is the
in a RED stack, a theoretical model was used. The relationship thickness of the diluted compartment (m), Kc is the concentrated
between the salt concentration (activity) in the two compartments (0.5 M NaCl) compartment conductivity (4.648 S/m) and Kd is the
(diluted and concentrated), the process temperature and the selec- diluted (0.05 M NaCl) compartment conductivity (0.551 S/m).
Eq. (6) predicts the theoretical membrane stack power output
tivity of the membrane is shown in the following equation:
and relates the maximum power output of the system to the
individual membrane characteristics, it can be used as a tool to compare different anion and cation exchange membranes with respect
to their performance in a RED stack. If Eq. (6) is used to predict
where Vo is the open circuit potential of the membrane stack (V), av the power output obtainable with only a cation exchange memis the average membrane permselectivity of an anion and a cation brane, then the average membrane pair permselectivity (av) is
exchange membrane pair, N is the number of membrane pairs, replaced by the individual membrane selectivity (). The resistance
R is the gas constant (8.314J/(mol K)), T is the absolute tempera- of the anion exchange membrane term will be removed in that
ture (K), z is the electrochemical valence, F is the Faraday constant case and the concentrated and diluted compartment thickness (d)
will be divided by two. The same procedure is applicable for anion
(96,485 C/mol), ac is the activity of the concentrated salt solution
exchange membranes. This calculated power output for only one
(mol/l) and ad is the activity of the diluted salt solution (mol/l).
type of the membrane is used to compare the different anion or
The stack resistance can be defined as the sum of the resis- cation exchange membranes with each other.
tances of the individual stack components as shown in the following
In the model a few assumptions were made: (i) concentration
equation:
polarization phenomena near the membrane surface are negligible
due to the small current densities obtained through the membranes. This phenomenon can be minimized with well designed
fluid dynamics between the membranes, (ii) the resistance of the
where N is the number of membrane pairs, A is the effective memelectrodes is assumed to be negligible compared to the membrane
brane area (m2),Raem is the anion exchange membrane resistance
resistance. This assumption is allowed when a large number of
2
2
(2 m ), Rcem is the cation exchange membrane resistance (2 m ), membrane cell pairs is used, (iii) the feed solution does not change
dc is the thickness of the concentrated compartment (m), dd is the
in concentration along the channels. This assumption has a strong
thickness of the diluted compartment (m), Kc is the concentrated
relationship with the feed channel design and in a well-designed
compartment conductivity (S/m), Kd is the diluted compartment system this effect can be neglected.
conductivity (S/m) and Rel is the electrode resistance (2).
In order to compare commercially available membranes with
The final stack power output of the RED stack can be found from
each other it is more desirable to convert the power output into
Kirchhoffs law and is defined as
power density by normalizing it for the membrane area (W/m2).
Power density per square meter of membrane can be easily calculated from the maximum power output:

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
270

where wmax is the maximum power density (W/m2), Wmax is

maximum power output (W),A is the effective membrane area (m2)
and Nm is the number of membranes.
Obtained power output per type the membrane (anion or cation
exchange membrane) is used only to benchmark membranes with
each other.
5. Conclusion

This paper evaluates the potential of commercially available

anion and cation exchange membranes for application in RED.
Different membrane properties and characterization methods are
discussed and a theoretical membrane model for RED was used to
allow fair comparison of the characterization results for application
in RED. Literature shows that the charge density has a strong
influence on the permselectivity and the membrane resistance, but
that there does not exist a straight forward relationship between
these properties. In general cation exchange membranes have a
higher charge density and a corresponding higher permselectivity
than anion exchange membranes. Heterogeneous ion exchange
membranes have a relatively low charge density compared to
homogeneous membranes and a much higher resistance due to the
membrane structure and the separation of charged domains in an
uncharged polymer matrix. This makes RED a potentially
attractive alternative for energy production.

REFERENCES
[1]. R.E. Lacey, Energy by reverse electrodialysis, Ocean Eng., 7

(1980) 147
[2]. J. Weinstein and B. Leitz, Electric power from differences
in salinity: the dialytic battery, Science, 191 (1976) 557559.
[3]. [Link]
[4]. M. Turek, Cost-effective electrodialytic seawater
desalination, Desalination, 153 (2002) 371376.
[5]. Piotr Dugocki, Kitty Nymeijer, Sybrand Metz and
Matthias Wessling, Current status of ion exchange
membranes for power generation from salinity gradients,
Journal of Membrane science, 319,July 2008, Pages 214222 .

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
271

Microbial fuel cells to produce electricity

Sameer Deshmukh, Yogesh Khambia
Wroclaw University of Technology
Wroclaw, Poland

Abstract- Microbial fuel cell (MFC) provides a new

opportunity for the sustainable production of energy in the form
of direct electricity from biodegradable compounds present in the
wastewater achieving the simultaneous water treatment. MFC is a
device that converts chemical energy to electrical energy with the
aid of the catalytic reaction of microorganism. This article
provides a brief introduction to MFC along with its envisaged
applications as one of the class of biological fuel cell.
Keywords: Microbial Fuel cell, Electricity

I.

INTRODUCTION

The high energy intensive conventional wastewater treatment

systems invites for the development of alternative treatment technology. There is a tremendous need to develop suitable and reliable
technology for the treatment of wastewater. Such treatment technology should be cost effective, requiring less energy for its efficient
operation and should generate less sludge. In addition, the treatment
system should recover energy to make overall operation of
wastewater treatment self sustainable.
Microbial fuel cell (MFC) provides new opportunity for the sustainable production of energy, in the form of direct electricity from
biodegradable compounds present in the wastewater, achieving
simultaneous wastewater treatment. MFC is a device that converts
chemical energy to electrical energy with the aid of the catalytic
reaction of microorganisms1. In a MFC, substrate (organic matter or
biomass) is oxidized in the anode chamber producing carbon
dioxide, protons and electrons2. Microorganisms here fulfill the role of
catalysts in analogs to chemical fuel cells. The electrons and the protons produced in the anode chamber ends up at the cathode, via
the external electrical circuit and through the proton exchange
membrane (PEM), respectively. In the cathode, an oxidant (normally oxygen) is being reduced. Eqs. (1)(3) illustrates the basic
reactions occurring in MFCs, in the case of a sucrose fed wastewater in anode and oxygen as electron acceptor in cathode.

Performance of a MFC is affected by the substrate conversion

rate, over-potentials at the anode and at the cathode, the PEM performance, and internal resistance of the cell2. The optimization of
MFCs requires extensive exploration of the operating parameters
that affect the power output.
All the literature review supports performance of MFC under
controlled conditions using different cultures. Enough attention
has not been paid so far on evaluating performance of MFC

exposed to changes in operating temperature and pH, which are

among the important aspects for field application of the MFCs.
After a brief outline of the long and winding history of
biological fuel cells, this article deals with the operating
principles of Microbial fuel cells and their technological
applications.
II. HISTORY OF BIOLOGICAL FUEL CELLS
For centuries, microorganisms, which transform food into an
electron flow, were only a biological curiosity; but now scientists
have made it possible to use them in watches and cameras
as power sources3. Luigi Galvani, who noticed twitching of
isolated frog leg when a brief electrical discharge was passed
through it4, was the first to observe the bioelectric
phenomenon as early as in 1790 and the term bioelectricity
was coined after that observation. In 1910, Michael Cresse
Potter, a professor of botany at the University of Durham, UK,

Figure1. A typical biological fuel cell representing current generation

with the help of microorganisms. The fuel generated by microbial metabolism
gets oxidized at the anode and usually oxygen is reduced at the cathode.

demonstrated that organisms could generate a voltage and

deliver current. Cohen at Cambridge revived Potters idea in
1931. He described how a batch of biological fuel cells
produced more than 35 V. But biological fuel cells became
popular in the 1960s, when the National Aeronautics and
Space Administration evinced interest in turning organic waste
into electricity on its long-haul space flights. Algae and bacteria were among the first organisms used in biological fuel cells.
During the mid 19th century, for the first time, Rohrback et al5
designed a biological fuel cell in which Clostridium butyricum
was used as a biological material to generate hydrogen by

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
272

glucose fermentation. In 1963, biological fuel cells were

already commercially available for use as a power source
in radios, signal lights and other appliances at sea.
However, these fuel cells were not a commercial success and
soon disappeared from the market. With the successful
development of technical alternatives, e.g. solar photovoltaics
for the energy supply on space flights, biological fuel cells
suffered a short setback. Later, during the oil crisis of the 70s
and 80s, the interest in the development of biological fuel
cells was revived. In 1966, Williams6 showed that biological
fuel cells powered by rice husk produced 40 mA at 6 V. Rice
husk is a potential source of lignocellulose, which on
fermentation yields many useful enzymes and biofuels like
ethanol that could be used in biological fuel cells. In 1969,
Yao et al.7 showed that glucose could be used as a fuel in the
presence of platinum-black. Later, Karube et al8 reported
the generation of about 300 mA electric current from an
Anabaena spp.-based biological fuel cell, in which
phosphoric acid was used as the electrolyte. Bennetto and coworkers have made noteworthy contributions to biological fuel
cells. They have developed and demonstrated improved
biological fuel cells using various microorganisms and
mediator systems. They showed that the mediators could
enhance both the efficiency of electron-transfer and the
reaction rate. Recently, Chaudhuri and Lovely9 have reported
that a microorganism R. ferrireducens can recover an electron
from glucose oxidation in the presence of Fe3+ up to 83%
without a mediator. Presently, efforts are being expended to
improve the performance of mediator-less biological fuel
cells as well as on finding an effective route to wire the
microorganism to the electrode so as to promote the
efficiency of electron-transfer.
III. Microbial fuel cells
The use of microorganisms in biological fuel cells eliminates
the isolation of individual enzymes, thereby providing cheaper
substrates for biological fuel cells. Microorganisms can be
used in four ways for producing electrical energy:

(iv) In the fourth configuration, the metal-reducing

bacterium having cytochromes in its outer membrane
and the ability to communicate electrically with the
electrode surface directly result in a mediator-less
biological fuel cell.
A brief description of the aforesaid configurations is
given below:
Microbial-systems producing hydrogen as fuel for conventional fuel cells: Various bacteria and algae, e.g. Escherichia
coli, Enterobacter aero-genes, C. butyricum, Clostridium
acetobutylicum and Clostridium perfringens, are known to be
active for hydrogen production under anaerobic condition10, 11.
The most effective hydrogen-producing microorganism is C.
butyricum12. E. coli and Enterobacter aero-genes are facultative
anaerobes and ferment both glucose and lactose as a carbon
source to produce hydrogen.
The conversion of carbohydrate to hydrogen is achieved by
a multienzyme system. In bacteria, it involves the conversion
of glucose to 2 mol of NADH, a reduced form of coenzyme 1,
namely -nicotinamide adenine dinucleotide of the vitamin
niacin, and 2 mol of pyruvate formed by EmbdenMeyerhof
pathway. Pyruvate is then oxidized through a pyruvate
ferredoxin oxidoreductase producing acetyl-CoA, CO2 and
reduced ferredoxin. NADHferredoxin oxidoreductase oxidizes
NADH and reduces ferredoxin. The reduced ferredoxin is
reoxidized to form hydrogen by hydrogenase. As a result, 4
mol of H2 are produced from 1 mol of glucose under ideal
conditions as shown below. In practice, however, H2 yield is
only about 25% of the theoretical value13. Improvement in
H2 production is possible by genetic engineering techniques and
screening of new hydrogen-producing baceria.

(i) Microorganisms can produce electrochemically active

substances through fermentation or metabolism. For the
purpose of energy generation, fuels are produced in
separate reactors and transported to the anode of a
conventional fuel cell. Accordingly, in this configuration,
the microbial bioreactor is kept separated from the fuel
cell.
(ii) In the second configuration, the microbiological
fermentation process proceeds directly in the anodic
compartment of the fuel cell.
(iii) In the third configuration, electron-transfer mediators
shuttle electrons between the microbial biocatalytic
system and the electrode. The mediator molecules accept
electrons from the biological electron transport chain of
the microorganisms and transport them to the anode of
the biological fuel cell.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
273

A H 2 /O 2 fuel cell comprising a platinum-blacknickel

mesh anode and a palladium-blacknickel mesh cathode, and
separated by a nylon filter operating at room temperature was
connected to a bioreactor producing hydrogen14,15. The
current and voltage output were dependent on the rate of
hydrogen production in the fermenter. For example, an opencircuit voltage (Voc) of 0.95 V and a short-circuit current
density ('sc) of 40 mA/cm2 were obtained at a H2 flow of 40
ml/min. The biological fuel cell operating under steady-state
conditions for a week produced a continuous current between
500 and 550 mA.
The immobilization of the biocatalyst is primarily important
in a bioreactor. In the biological fuel cell configuration described
above, the immobilization of hydrogen-producing bacteria, C.
butyricum is of great value as this stabilizes the relatively
unstable hydrogenase system. The immobilization steps could
be the trapping of microorganisms into polymeric matrices
of polyacrylamide, agar gel or filter paper. The
immobilized microbial cells continuously produced H2
under anaerobic conditions for several weeks, whereas nonimmobilized bacterial cells were fully deactivated in less than
two days.
Microbial systems producing electrochemically active
metabolites in the anodic compartment of biological fuel cells:
In this configuration, the fermentation process is conducted
directly at the electrode surface supplying the anode with H2
fuel. Additional by-products of the fermentation process,
namely formic acid, acetic acid and lactic acid are also utilized
as fuels16. Besides, the base substrate glucose used for the
fermentation process, byproducts could as well contribute to the
anodic current. Hence, H2 provided by the microorganisms can
be a source of anodic current as indicated by the side reactions
below.

In addition to fuels like H2, formic acid, lactic acid and

sulphur-containing electrochemically active metabolites like
S2 species, can be produced during the fermentation of
lactate by Desulfovibrio desulfuricans, which are known to be
sulphate-reducing bacteria, shown as follows.

with iron-containing proteins, e.g. cytochromes, causing the

electron transport systems to be blocked. S 2- species poison
many metallic electrodes because of their strong and
irreversible adsorption. In a typical experiment, porous graphite
electrode impregnated with cobalt hydroxide as catalyst was
used as the anode. Cobalt hydroxide undergoes transition into a
highly catalytically active cobalt oxide/cobalt sulphide mixture.
A biological fuel cell has been constructed with the above
biocatalytic anode along with a graphite cathode activated with
iron(II) phthalocyanins and vanadium(V) compounds. The
anode and cathode are separated to maintain anaerobic
conditions at the anode compartment. The poor performance
of fuel cells arising due to the adsorption of by-products on
the electrode surface is improved by an electrode modification
process. The microorganisms processing glucose in a tank
of the fermentation fluid are continuously pumped through
a separate anode space. This is separated from the cathode
space by a semi-permeable membrane. These bio-fuel cells
have a new type of anode where a platinum electrode or a
platinized graphite electrode is coated with a layer of the
electrically conducting polyaniline, which is both
biocompatible and electro-catalytically active. It absorbs
electrons from the metabolism of the bacteria and transfers
them to the anode. In this way, it plays a decisive role in
current flow. During operation, the bacterial metabolic
products along with the by-products of the electro-catalytic
oxidation process settle on an uncoated anode and rapidly
deactivate it. The polymer slows down this process considerably.
Additional regular voltage pulses chemically convert the
deposit and release them from the anode surface. This fuel cell
continuously provides up to 1.5 mA/cm2 of current17
Mediator-coupled microbial fuel cells: Reductive species
generated by metabolic processes inside microbial cells are
isolated by a microbial membrane. Thus, the contact of the
microbial cells with an electrode usually results in only a
diminutive electron-transfer across the membrane of the
microbes18, except in some special cases as described later.
The electro-active groups responsible for the redox activity of
enzymes present in the microbial cells are deeply buried
inside their prosthetic groups, which leads to poor
electrical communication between the cells and the
electrode surface. The cells can, however, be wired to the
electrode surface with the help of mediators. Low molecular
weight redox species may assist the shuttling of electrons
between the intracellular bacterial space and an electrode,
and are referred to as mediators. The working principle of
these mediators is shown schematically in Figure 2.
The mediator molecules should meet the following requirements19:
(i)

The presence of sulphides in the medium results in the inhibition

of the metabolic bacterial processes because of their interaction

The oxidized mediator should easily penetrate through the

bacterial membrane to reach the reductive species inside
the bacteria.
(ii) The redox potential of the mediator should match the

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
274

potential of the reductive metabolite.

(iii) None of the oxidation states of the mediator should
interfere with other metabolic processes.
(iv) The reduced mediator should easily escape from the cell
through the bacterial membrane.
(v) Both the oxidized and reduced states of the mediator should
be chemically stable in the electrolyte solution, should
be easily soluble, and should not adsorb on the bacterial
cells or electrode surface.
(vi) The electrochemical kinetics of the oxidation process for
the mediator-reduced state at the electrode should be fast.
A variety of organic compounds have been studied in
combination with bacteria to test the electron-transfer
efficiency of the mediator between the microorganism and
anode surface.

presence of E. coli immobilized on graphite particles. Table

2 summarizes the characteristics of biological fuel cells
containing mediators.23
The electrodes should be designed so as to facilitate
electrical contact between a biocatalytic system and an
anode, and to improve the cell output. The mediators can be
coupled to the microorganisms in three ways1 (Figure 3): (i)
as diffusional mediator shuttling between the microbial
suspension and the anode surface, (ii) a diffusional mediator
shuttling between the anode and microbial cells covalently
linked to the electrode. The microbial cells can be covalently
linked to the electrode surface having COOH groups through
amino groups of the microbial membrane resulting in the
formation of amide bond. Standard organic reagents like
carbodiimide and acetyl chloride are used to link the microbial
cells to the surface, and (iii) mediator adsorbed on the
microbial cells providing electron transport from the cells to
the anode.
Mediator-less microbial fuel cells23: Fe(III)-reducing microorganisms are found to be electrochemically active as they
have cytochromes in their outer membranes. It was first
demonstrated with the Fe(III) reducers, Shewanella
putrefaciens, that these can be used as a catalyst in a mediatorless microbial fuel cell. Recent studies have demonstrated that
Fe(III) reducing microorganisms of the family Geobacteraceae
can directly transfer electrons on to electrodes. However, the
range of electron donors that these organisms can use is limited
to simple organic acids such as acetate.

Thionine and organic dyes20 have been frequently used as

mediators in biological fuel cells. Redox potentials and structures
of some of the electron relays (mediators) are summarized in
Table 1. The overall efficiency of the electron-transfer mediators
depends on many parameters, particularly on the
electrochemical rate-constant of mediator re-oxidation, which
depends on the electrode materials21. It is difficult to realize the
perfect conditions for electron transport from a bacterial
cell to an electrode. A mixture of two mediators can be useful in
optimizing the efficiency. Two mediators, namely thionine and
Fe(III) EDTA, are employed with the biocatalyst E. coli for
the oxidation of glucose22. It was found that thionine is
reduced over 100 times faster than Fe(III) EDTA. But the
electrochemical oxidation of thionine is much slower than
the oxidation of Fe(II) EDTA. Therefore, electrons obtained
from the biocatalysed oxidation of glucose are transferred
mainly to thionine. The reduced thionine is rapidly reoxidized
by Fe(III) EDTA. The reduced Fe(II) EDTA transfers the
electrons to the anode, and is kinetically fast.
Another example of a mixture of mediators promoting
electron-transfer across the anode is demonstrated with the help
of methyl viologen and 2-hydroxy-1,4-napthoquinone in the
case of bioelectrocatalysed oxidation of glucose in the

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
275

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
276

the presence of Fe(III). The stoichiometry of glucose utilization and Fe(III) reduction can be explained as below.

The performance was tested in a system that comprised a reactor

containing an anaerobically growing suspension of E. coli
K12 in a standard glucose medium (55 mM glucose) and
the fuel cell consisting of an anode compartment through
which the bacterial medium was pumped. The cathode was
woven graphite and the catholyte was a 50 mM
ferricyanide solution in a phosphate buffer, akin to the
buffer in a bacterial medium.
The recovery of electrons from glucose oxidation is 83%
of the theoretical value available from glucose oxidation.
Microbial growth is supported by energy derived from the
electron-transfer process itself and results in a stable, longterm power production. This type of microbial fuel cell
exhibits many of the desirable features of secondary
batteries, including the ability to be recharged to the nearly
original charged state subsequent to the discharge, the ability
to accept fast recharge, reasonable cycle-life and low
capacity-loss under open-circuit conditions as well as on
prolonged storage under idle conditions. Thus, mediator-less
fuel cells have an advantage over those with mediators in
terms of cost as well as non-desirability of toxic mediators. In
mediator-less fuel cells, there is also ample room to increase
the efficiency of electron-transfer.
Comparision Between Chemical fuel cell and biological fuel
cell.
Metal-reducing bacteria are the most used species in this
type of fuel cells. Studies have been conducted to understand
the mechanism of Fe(III) reduction by bacteria such as
Shewanella putrefaciens and Geobacter metallireduce.
Development of mediator-less microbial fuel cells using
Shewanella putrefaciens IR-1 has been reported by Kim et
al.48. In these fuel cells, the electrochemical activity of the
microorganism has been confirmed by cyclic voltammetry when
grown under anaerobic conditions, although no activity was
found when they were grown under aerobic conditions. A
gradual decrease in both coulombic yield and maximum
current values was observed during the sequential batch
operation of microbial fuel cells. Recently, there is a report on
the bacteria, Rhodoferax ferrireducens that can be used in
microbial fuel cells effectively without a mediator22 . R.
ferrireducens was isolated from anoxic sub-surface sediments
of Oyster Bay, Virginia, USA as a dissimilatory Fe(III)
reducing microorganism. R. ferrireducens grows on glucose in

III. Applications of biological fuel cells [23]

In about 200 years from now, vehicles will not have petrol
tanks because our petroleum reserves will be depleted. An
alternative, which is less wasteful and cleaner, is to power

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
277

vehicles directly with carbohydrates using biological fuel

cells. The energy liberated during the complete oxidation of a
monosaccharide like glucose or a disaccharide such as sucrose
to carbon dioxide and water is about 16 x 106 J/kg, which is
about 5 kWh of electrical energy and is just less than half
the energy that can be obtained from equivalent amounts of
fuels such as octane. But the efficiency of burning
carbohydrate in biological fuel cells is potentially greater
than burning gasoline. For example, a medium-sized car that
needs about 200 Wh/km, could travel 2530 km on one
kilogram of a concentrated solution of a carbohydrate.
Accordingly, 50 l of a strong sugar solution would give the
car a range of more than 1000 km.

of California that a miniaturized microbial fuel cell could be

integrated with MEMS based implantable medical devices. In
comparison to lithium batteries as power sources for implantable
devices, microbial fuel cells are smaller, less expensive and have
a longer shelf-life.
A mobile robot platform has been built that derives its
propulsive power from the digestion of real food in a bioreactor.
This robot called Gastrobot used harmless yeast and
adopted a gas propulsion system without an electrical
generation. When a carbohydrate-rich aqueous solution is mixed
with yeast in a reaction chamber, fermentation takes place with
liberation of carbon dioxide and it leads to substantial pressure
build-up when constrained. Energy available in this
pressurized CO2 gas is used to propel the Gastrobot, which is
achieved purely by mechanical means.

Slugbot is a robot that utilizes the electrical power

produced from biomass (Figure 10). Slugs are common
pests found on agricultural land in the UK, where farmers spend
about 20 million per annum on buying and spreading
molluscicides. Slugbot ferments slug mass and converts it into
electrical energy, which it uses to catch the slugs in the
field. Stuart Wilkinson at the University of South Florida,
who is currently interested in the development of hybrid
biological fuel cells, has developed a meat-eating robot Chew
Chew, which is powered by a microbial fuel cell and uses meat
as a fuel. This has three wagons, each about a metre long.
These robots could even kill human beings accidentally.
One could construct molecular biological cells, which could
deliver electrical energy to remove tumours and cancerous
cells, and could act as drug-delivery systems.

A biological fuel cell that is just 0.07 cm2 in area, has been
designed to generate as much as 300 tV for 2 h, an amount
sufficient to operate tiny devices, including microscopic drugdelivery systems. Such a microbial fuel cell could power
implantable medical devices as shown in Figure 9 and
could help individuals who require regular doses of drugs,
for example, AIDS patients. Apart from its small size, the
system is unique because it utilizes glucose, a sugar present in
the blood stream, as fuel. Such miniaturization has become
possible due to the advent of BIOMEMS (Bio-micro
electromechanical systems). Since biological fuel cells are
known to supply low but stable power density, they are
suitable for running devices constructed using MEMS
technology. It has been shown by researchers at the University

Another important application of microbial fuel cells is in

the field of waste water engineering. Microorganisms can
discharge the dual duty of degrading effluent and generating
power. When microorganisms oxidize organic compounds
present in waste water, electrons are released yielding a steady
source of electrical current. If power generation in these
systems can be increased, microbial fuel cells may provide a
new method to offset operating costs of waste water
treatment plant, making advanced waste water treatment
more affordable in both developing and industrialized nations.
Different designs of fuel cell reactors based on chemical
engineering principles like fluidized bed reactors, packed
bed reactors, etc. are under trial. Tests have been
conducted using single chamber microbial fuel cell
(SCMFC) containing eight graphite electrodes as anodes and
a single air cathode. The system was operated under

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
278

continuous-flow conditions with waste water. The prototype

SCMFC reactor generated an electric power of 26 mW/m2,
while removing up to 80% of chemical oxygen demand of the
waste water.23

Conclusions
The main challenge is in the electrical coupling of the
biological component of the system with the electrodes of
the fuel cells. By tapping the complete multi-enzyme
metabolic pathways inside living cells, microbial fuel
cells could last long and could utilize complex biofuels. A
detailed characterization of the interfacial electron-transfer
rates, biocatalytic rate constants and cell resistance is
essential for the construction of microbial fuel cells.
Although microbial fuel cells are in their infancy, the prospects
of their applications look attractive.

REFERENCES
[1]. Allen, R.M., Bennetto, H.P., 1993. Microbial fuel cells: electricity production
from carbohydrates. Appl. Biochem. Biotechnol., 2740
[2]. Rabaey, K., Verstraete, W., 2005. Microbial fuel cells: Novel
biotechnology for energy generation. TRENDS in Biotechnol. 23
(6), 291298.
[3]. Bennetto, H. P., Mason, J. R., Stirling, J. L. and Thurston, C. F.,
Research and development in non-mechanical electrical power
sources. In Power Sources 11 (ed. Pearce, L. J.), 1987, p. 373.
[4].Cole, K. S., In Membranes, Ions and Impulses, University of California Press, 1972.
[5]. Rohrback, G. H., Scott, W. R. and Canfield, J. H., Biochemical
fuel cells. In Proceedings of the 16th Annual Power
Sources Conference, 1962, p. 18.
[6]. Williams K. R., In An Introduction to Fuel Cells, Elsevier, Amsterdam, 1966, p. 248.
[7]. Yao, S. J., Appleby, A. J., Geise, A., Cash, H. R. and Wolfson,
S. K., Anodic oxidation of carbohydrates and their derivatives in
neutral saline solution. Nature, 1969, 224, 921922.

H., Photochemical energy conversion using immobilized blue-green

algae. J. Biotechnol., 1986, 4, 7380.
[9]. Chaudhuri, S. K. and Lovely, D. R., Electricity generation by
direct oxidation of glucose in mediator-less microbial fuel cells. Nature
Biotechnol., 2003, 21, 12291232.
[10]. Gottschalk, G., In Bacterial Metabolism, Springer Verlag, New York,
1979, 2nd edn.
[11]. Heyndrickx, M., Devos, P. and De Ley, J. U., Hydrogen production
from chemostat fermentation of glucose by Clostridium butyricum and
Clostridium pasteurianum in ammonium and phosphate limitation.
Biotechnol. Lett., 1990, 12, 731736.
[12]. Suzuki, S. and Karube, I., Energy production with immobilized
cells. Appl. Biochem. Bioeng., 1983, 4, 2813 10.
[13]. Suzuki, S., Karube, I., Matsuoka, H., Ueyama, S., Kawakubo, H.,
Isoda, S. and Murahashi, T., Biochemical energy conversion by
immobilized whole cells. Ann. N.Y. Acad. Sci., 1983, 413, 133 143.
[14]. Karube, I., Suzuki, S., Matsunaga, T. and Kuriyama, S., Biochemical energy conversion by immobilized whole cells. Ann. N.Y.
Acad. Sci., 1981, 369, 9198
[15]. Suzuki, S., Karube, I., Matsunaga, T., Kuriyama, S., Suzuki, N., Shirogomi,
T. and Takamura, T., Biochemical energy conversion using immobilized
whole cells of Clostridium butyricum. Biochimie, 1980, 62, 353358.
[16].Karube, I., Matsunaga, T., Tsuru, T. and Suzuki, S., Biochemical fuel
cell utilizing immobilized cells of Clostridium butyricum. Biotechnol.
Bioeng., 1977, 19, 17271733.
[17]. Schroder, U., Nie er, J. and Scholz, F., A generation of microbial fuel
cells with current output boosted by more than one order of magnitude.
Angew. Chem., 2003, 115, 29862989.
[18]. Allen, M. J., Biofuel cells. In Methods in Microbiology (eds Norris, J.
R. and Ribbon, D. W.), Academic Press, New York, 1972, pp. 247283
[19]. Wilkinson, S., Gastrobots benefits and challenges of microbial fuel
cells in food powered robot applications. Autonomous Robots, 2000, 9, 99
111.
[20]. Ikeda, T., Kato, K., Tatsumi, H. and Kano, K., Mediated catalytic
current for the oxidation of ethanol produced by Acetobacter aceti cells
suspended in solution. J. Electroanal. Chem., 1997, 440, 265269.
[21]. Ikeda, T., Kato, K., Tatsumi, H. and Kano, K., Mediated catalytic
current for the oxidation of ethanol produced by Acetobacter aceti cells
suspended in solution. J. Electroanal. Chem., 1997, 440, 265269.
[22]. Tanaka, K., Vega, C. A. and Tamamushi, R., Thionine and ferric chelate
compounds as coupled mediators in microbial fuel cells. Bioelectrochem. Bioeng.
1983, 11, 289297.
[23]. Shukla A. K. ; Suresh P.; Berchmans S. ; Rajendran A, Biological fuel
cells and their applications, Current science, 2004, vol. 87, pp. 455-468.

[8]. Karube, I., Ikemoto, H., Kajiwara, K., Tamiya, E. and Matsuok,

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
279

Cable Shielding to Minimize Electromagnetic

Interference
Anke Frbel
Cottbus University of Technology, Germany
Wrocaw University of Technology, Poland

Abstract - A cable shield is necessary to prevent emission of

electromagnetic waves from the cable respectively to protect data
and signal conductors from external electromagnetic interference
(EMI). The effectiveness of a cable shield installation depends on
the kind of EMI to be shielded and the type of termination at both
ends. This paper depicts the different types of cable shielding.
Moreover shielding effectiveness is analyzed.

I.

INTRODUCTION

Electromagnetic shielding reduces, or rather prevents

coupling of undesired radiated electromagnetic energy in
electrical equipment. It is established over a large part of the
electromagnetic spectrum from DC to microwave frequencies.
In general, shielding is produced by insertion of a metallic
barrier in the path of electromagnetic waves between the
source of the radiation and the device which is supposed to be
protected. The shielding may be applied at the source (if the
source is known) or at the susceptible equipment. Fig.1.
illustrates the two modes of shielding.
Cable shielding is commonly used for data and signal cables.
The design of cable shielding and grounding is important for
the reduction of electromagnetic interference. Each application
requires individual considerations given that parameters such
as cable lengths, noise frequency, signal frequency and cable
termination methodology impact the final result. Improperly
cable shielding can actually increase noise coupling and thus
make the problem worse.

SHIELDING THE
NOISE SOURCE

NOISE
SOURCE

SHIELDING
SUSCEPTIBLE
EQUIPMENT

NOISE
SOURCE

SUSCEPTIBLE
EQUIPMENT

SUSCEPTIBLE
EQUIPMENT

Figure 1. Modes of shielding: Shielding the source of noise

or shielding the susceptible equipment.

II. TYPES OF SHIELDED CABLES

Since data or signal cables carry broadband signals, up to
high frequency ranges, they need to be shielded to minimize

radiated coupling. In addition, the impedance has to be

controlled. The selection of a certain type of cable as well as
the arrangement of signals and grounds rather shields in the
cable determine the characteristic impedance of the
transmission path.
In the following paragraphs the various versions of shielded
cables are introduced.
A. Coaxial cables
A coaxial cable consists of an inner conductor, surrounded
by an insulating layer which is then surrounded by another
conductive layer, typically a fine woven braid for flexibility or
a thin metallic foil, and than covered again with a thin
insulating layer on the outside, shown in Fig. 2. The shield is
grounded at multiple points at high frequencies and a single
point at low frequencies. Coaxial cables are used in a
frequency range of 20 kHz up to 50 GHz.

Figure 2. Coaxial cable cutaway

B. Triaxial cables
A triaxial cable is similar to coaxial cables but with another
shield isolated from the signal return shield. The second shield
is grounded, as shown in Fig. 3. Thus, triaxial cables provide a
greater rejection of interference than coax.
C. Twinaxial cables
A twinaxial cable is a two-wire twisted balance line in a
ground shielding braid, as shown in Fig. 4. The twisting
provides cancellation of any induced noise voltage pickup
caused by a leakage of the low-frequency magnetic field
through the copper braid. Twinax are used up to 10 MHz.
D. Quadraxial cables
A quadraxial cable is a double-shielded twinax, as shown in
Fig. 5. The outer shielding is grounded to the earth while the
inner shielding is connected to the system ground. If there is no
system ground available, both shields are earth grounded.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
280

2.

Magnetic field coupling, the incident wave is polarized

normal to the loop formed by the cable and the ground
plane.
The EMI voltage pickup in a cable increases with frequency.
At high frequencies resonance phenomena produce maximum
induced voltages for a certain cable length l such that

Figure 3. Shielded triaxial cable

Figure 4. Shielded twinaxial cable

Figure 5. Shielded quadraxial cable

both ends grounded + H-field excitation

no resonance
both ends grounded + E-field excitation
resonance for l = k / 2
one end grounded + H-field excitation
resonance for l = (2k + 1) / 4
one end grounded + E-field excitation
resonance for l = (2k + 1) / 4

At low frequencies for E-field excitation it is more efficient

to ground both ends, whereas for H-field excitation one end
grounding has to be favored, since this eliminates the formation
of a current loop by the cable and the ground plane.
At high frequencies both ends grounded configurations
avoids resonances for E-field and H-field excitations.
In practice one ground connection is often preferred, since
this avoids ground loops. However, for short cables, at low
frequencies, the voltages induced by EMI at both ends of a
coaxial cable become nearly equal and one end grounding is
needed for E-field as well as for H-field excitations.

III. CABLE SHIELD GROUNDING

If a shielded cable is used to connect two systems, the shield
has to be connected to a single ground reference. In order to
prevent that electromagnetic energy penetrates through the
shield, the outer surface of the shield has to be grounded, as
shown in Fig. 6.
It is possible to ground the shield at one end (asymmetric), at
both ends (symmetric) or in intervals along the length of the
cable. The effectiveness of these different methods depends on
the electromagnetic coupling mode and the length (l / ) of the
cable used for the interconnection.

Figure 6. Cable grounding

There are two kinds of electromagnetic coupling in a cable:

1. Electric field coupling, the incident wave is polarized
parallel to the conductor length, and

IV. TRANSFER IMPEDANCE OF CABLE SHIELD

The finite conductivity of shielding material as well as its
small thickness and small openings in the braid allow
electromagnetic fields to penetrate through the shield and
induce currents in the line. Therefore it is necessary to evaluate
the ultimately shielding effectiveness.
It is difficult to measure the field inside of a cable shield
accurately. Moreover, the voltage measured at the end of the
line depends on the type of termination. Therefore a definition
of shielding effectiveness using the ratio of fields before and
after the shield, or the ratio of voltage induced with and
without a shield, is not adequate. An evaluation of the shielding
effectiveness regarding the transfer impedance is more
common.
The transfer impedance of a cable shield relates the current
I S flowing on the shield surface to the longitudinal induced
voltage Vi per unit length on the outer side of this surface (see
Fig. 7).
The sheath current I S may result from an externally incident
field or ground potential difference between the two ends of the
cable.
Hence
Vi = I S Z t

where Z t is the transfer impedance of the shielded cable.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
281

(1)

Zt =

(1 + j )t /
1
2at sinh[(1 + j )t / ]

(3)

where a is the inner radius of the shield, t is its wall

thickness, is the conductivity of the shield and is the skin
depth of the shield.
At low frequencies and t / << 1 the transfer impedance is
Zt =

Figure 7. Model of transfer impedance coupling in coaxial cable

Z t is expressed in ohms normalized to a one-meter shield

length. The better the shielding the lower are the values of Z t .
At frequencies below 100 kHz, Z t is practically equal to the
DC shield resistance RDC . Above several MHz a capacitive

[1]

Z t = Z r + jLt + (1 + j )Ls

[2]

The transfer impedance of a single braided cable can be

reduced by wrapping a metallic band (e.g. aluminum), or rather
a conductive envelope (e.g. polycarbonate) over the braid. This
way the coupling inductance Lt is reduced.
For thin-walled tubular shields the transfer impedance can be
expressed as follows:

(4)

as mentioned above.
The transfer impedance and in order to that the shielding
effectiveness differs significantly for the different types of
shielded cables. Most common in use are solid semi-rigid
coaxial, braided triaxial, braided coaxial, shielded quadraxial
and shielded twinaxial cables (with descending order of
effectiveness).

coupling between the shield and the inner conductor has to be

considered. At frequencies above 10 MHz, Z t is proportional
to the leakage inductance.
A single braided shield is a solid tube with rhombic or
elliptical holes. The transfer impedance consists of three
components. One of them is the diffusion impedance Z r which
expresses the relation between the shield current and the
longitudinal electric field caused by the finite conductivity of
the equivalent shield tube. The second component is the
coupling inductance Lt term. It considers the magnetic field
coupling through the openings in braided wire shields and the
inductance between two interwoven parts of the braid. The
third component is the so called skin inductance Ls , which
results from magnetic fields penetrating the shield.
(2)

1
= RDC
2at

V. CONCLUSIONS
This paper presented analytical and practical aspects of cable
shielding to mitigate electromagnetic interference. It was
shown that the best shielding for any application depends on
the application itself. The best solution for one situation may
be bad for another set of conditions.
REFERENCES

[3]
[4]

[Link] Kodali, Engineering electromagnetic compatibility: principles,

measurements, and technologies, vol. I. New York: Institute of Electrical
and Electronics Engineers, 1996, pp. 199-227.
[Link], Power Quality, vol. I Florida: CRC Press LLC, 2002, pp.
154-164.
G. Vijayaraghavan, Mark Brown, and Malcom Barnes, Practical
Grounding, Bonding, Shielding and Surge Protection, vol. I Oxford:
Newnes, 2004
[Link] February 7, 2009

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
282

Fault Location System for Transmission Lines in One-terminal

By using Impedance-Traveling Wave Assembled Algorithm
Sachin R. Yerekar

Department of Electrical Engineering

Veermata Jijabai Technogical Institute,
Matunga, Mumbai-400019
sachin_yerekar@[Link]
Abstract: - Considering roundly reliability and accuracy of
one terminal fault location system for transmission lines, an
impedance-traveling wave assembled algorithm, which combines
measurement impedance method with traveling wave method, is
presented. At first, it uses measurement impedance method to
calculate roughly fault distance. Then regions of reflected waves
from fault point and opposite bus to detective bus are confirmed
respectively through the distance. At last, exact time of reflected
wave from fault point to detective bus and that of reflected wave
from opposite bus to detective bus are identified respectively in
corresponding regions, and fault location is implemented
according to the result detected. It has complementarities of both
methods, because measurement impedance method guarantees
reliability and traveling wave method improves accuracy. Results
of simulation are that located errors are within 100m if fault
location based on traveling wave algorithm is available. It is
improved that assembled algorithm proposed is correct.

is presented. The assemble algorithm have complementarities,

because measurement impedance method guarantees reliability
and traveling wave method improves accuracy [7], [8].
[Link] - TRAVELING WAVE ASSEMBLED
ALGORITHM:
The algorithm proposed includes two parts, one is
measurement impedance method, and the other is traveling
wave method.
A. Measurement impedance method:
An algorithm named measurement impedance method is
proposed in document [9]. It eliminates fault resistance for
calculating fault distance according to measurement
impedance, transmission line positive-sequence impedance
angle and detective currents.

Keywords:- Accuracy; Assembled algorithm; Fault currents;

Fault location; Fault resistance; Measurement impedance; Oneterminal; Reliability; Transmission lines; Traveling wave.

I. INTRODUCTION:
Transmission lines are vital in electric power systems,
Fig.1 Fault diagram of three-phase power system
because they shoulder missions of transmitting power energy.
However, they are also the parts which are bounded to have
According to fault location algorithm based on
faults, and it is very difficult to find the faults. Thus, fast and
measurement
impedance method, fault distance can be
accurate fault location for transmission lines, which is an
calculated
by
Equation
(1).
imperative problem to be solved, plays an increasingly
important role in electric power systems. It not only saves a
large amount of manpower and material resource, but also
(1)
makes it easy to find some covert faults, such as the insulator
Where
a
=
Re[I
/
I
]
,
b
=
I
[I
/
I
]
,
x
is
fault
distance
Mf
M
m Mf
M
puncture. Various fault location methods have been developed
from
bus
M
to
fault
point
F,
x
is
positive-sequence
reactance
1
and fault location techniques have been a subject for many
per kilometer of transmission lines, L is transmissions line
researchers [1].
One-terminal fault location system just installs fault positive-sequence impedance angle, IMf is fault component
location device on one terminal of transmission lines, and current, UM and IM are respectively detective voltage and
doesnt have to use communication device, so it is more current with the definition of distance protective measurement
economical than two-terminal fault location system [2], [3]. In impedance, thats RM + jXM =UM / IM .
one-terminal fault location system, there are two categories,
In the algorithm named measurement impedance method,
one is based on power frequency quantities, and the other is its assumed that the phase angle between fault current If and
based on traveling wave. The former is economical and easy to fault component current IMf is zero, so the fault error will be
implement, especially, measurement impedance algorithm is inevitably led.
the most robust algorithm of one-terminal fault location based
Simplicity and reliability are the main characteristic of
on power frequency quantities [4], [5], but has no high algorithm based on measurement impedance. Its basic
accuracy. The latter has high accuracy except of reliability [6]. principle is the same as the one of distance protection. With
However, both reliability and accuracy are vital in fault measurement impedance, distance relay determines the fault
location technology. Thus, considering roundly reliability and section and trip fault line selectively. So the algorithm based
accuracy of one-terminal
fault location system, an impedance- on measurement
can be
boundary the fault
8. EEEIC International Conference on Environment and Electrical
Engineering, impedance
Karpacz, Poland,
Maytaken
10-13,to2009
traveling wave assembled algorithm, which combines 283 section of transmission lines [8].
measurement impedance method with traveling wave method,

B. Traveling wave method:

When a fault occurs in transmission lines, fault traveling
wave will transmit from fault point to detective bus and
opposite bus along the transmission lines. Because of
discontinuous impedance, fault traveling wave to detective bus
will be reflected at detective bus and transmit to fault point,
then will be reflected at fault point and transmit to detective
bus for the second time. For the same reason, fault traveling
wave to opposite bus will be reflected at opposite bus and
transmit to fault point, and may be refracted at fault point and
transmit to detective bus.

Where v' , t0 and x' are unknown, denoting respectively as the

true velocity of traveling wave, the time when fault occurs and
fault distance.
The equation as in (4) can be obtained from equation as in
(3).

(4)
In case of that t2 can be detected, but t3 cannot be,
simultaneous equation will be expressed as in (5).

Fig.2 Diagram of fault traveling wave propagation

As shown in Fig.2, L is the length of the fault lines, x is the

fault distance from detective bus M, t0 is the time when fault
occurs, t1 is the time when initial traveling wave arrives at
detective bus, t2 is the time when traveling wave reflected at
fault point arrives at detective bus, t3 is the time when
traveling wave reflected at opposite bus and refracted at fault
point arrives at detective bus.
A great deal of computer simulation showed that the
largest error of algorithm based on measurement impedance is
within the limit of 10% under different fault case [8]. So the
actual distance to fault point is sure in the region of
(x-10%L, x +10%L). tM , tN and t are defined as the needed
time that fault traveling wave transmits from fault point to
detective bus M, the needed time that fault traveling wave
transmits from fault point to opposite bus N, and the
corresponding time error of the distance error of 10%L .
They are expressed as in (2).

(2 )
Where v is defined as velocity of traveling wave which can be
calculated according to lines parameters or be obtained from
former velocity of traveling wave modified by (4).
Using t1 as time reference, t2 is in the region of
( t1+2tM -2 t , t1+2tM+2 t ) and t3 is in the region of
( t1+2tN -2 t , t1+2tN+2 t ).
In case of that both 2 t and 3 t can be detected,
simultaneous
equation will be expressed as in (3).

(5)
The equation as in (6) can be obtained from equation as in
(5).

(6)
In case of that t3 can be detected, but t2 cannot be,
simultaneous equation will be expressed as in (7).
(7)
The equation as in (8) can be obtained from equation as in

(7).

(8)
In worst case of that both t2 and t3 cannot be detected, t0
and x' will be expressed as in (9).

(9)

C. Analysis in theory:
Simplicity and reliability are the main characteristic of
algorithm based on measurement impedance. Its basic
principle is the same as the one of distance protection. With
measurement impedance, distance relay determines the fault
section and trip fault line selectively. So algorithm based on
measurement impedance can be taken to boundary the fault
section of transmission lines.
In the algorithm, exact time of reflected wave from fault
point to detective bus and that of reflected wave from opposite
bus to detective bus are identified respectively in the regions of
( t1+2tM -2t , t1+2tM+2t ) and ( t1+2tN -2t , t1+2tN+2t ). So
(3)
8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
t2 and t3 in full region of
284 its unnecessary to detect
[t1,t1+2L/v], but in its sub-regions of (t1+2tM -2t ,

t1+2tM+2t ) and
( t1+2tN -2 t , t1+2tN+2 t ). As a
result, total length of its sub-regions is v L t 8 = 0.8 , only
40% of the length of full region. Moreover, the algorithm can
also decrease the disturbance of traveling wave reflected from
neighbor buses.
In case of that both t2 and t3 can be detected, the precision
of fault location can avoid the effect of traveling wave velocity,
whats more, the velocity of traveling wave can be modified
online [11]. Because of the velocity of traveling wave modified
endlessly, it will approach the true velocity. In case of that just
t2 or t3 can be detected, the precision of fault location calculated
by Equation (6) or (8) is hardly affected by the effect of
traveling wave velocity. In worst case of that both t 2 and t 3
cannot be detected, fault location based on traveling wave
algorithm isnt available, but fault location algorithm based on
measurement impedance method can guarantees reliability.
III. SIMULATION TEST:
In order to demonstrate the correctness of the assembled
algorithm proposed, a model with length of 300km, voltage
grade of 500kV and frequency of 50Hz is built for simulation.
As shown in Fig.3, single transmission lines between bus M
and bus N connect Equivalent system M and Equivalent
system N. Fault distance can be set by the means of changing
the lengths of line part1 and line part2. Fault type and fault
resistance can be set through the module of Fault Setting. Ports
for measurement are set at bus M. Waveforms of currents and
voltages at bus M can be detected by the way of oscillograph
module, and data of voltages and currents will be put out
through the outputs of Out1 and Out2.
Transmission lines distributed parameters are given as
following:
r1 =0.0270 / km , x1 =0.2783 / km ,
c1 =0.0127 F / km , r0 =0.1948 / km ,
x0 =0.6494 / km, c0 =0.0090 F / km .
Parameters Equivalent system M and Equivalent system N
are given as following:
Z M1= ZN1 =2.534 + j 120.46 () ,
ZM0 = ZN0 =1.121+ j 40.23 ( ).
The velocity of traveling wave can be obtained from the
transmission lines distributed parameters [11]:
v=(1/l1c1)= 2.98137 105 km / s

Single phase-to-grounded fault with distance of 100km

from bus M is taken as an example. Setting sample frequency
at 2 kHz and sampling ten power frequency cycles, fault
distance of x=102.4km is calculated via Equation (1) with
Fourier transform, and results of tM=343.5s, tN=662.8 s
and
t =100.6 s are also calculated. Then just
keeping the current output of Out1 and setting sample
frequency at 2MHz and sampling half power frequency cycles,
data are treated with using digital filter, modal transform and
wavelet analysis [12],[13]. According to maximum likelihood
estimation, time reference of t1=2817 s is obtained, and
times of t2=3488 s and t3= 4158 s are identified
respectively in corresponding regions of (3302.8,3705.2) and
(3941.4, 4343.8) as shown in Fig.4. The results of t0 = 2481.5
s, v'= 2.98211105 km/ s and x'=100.05 km are calculated
finally.
Changing respectively fault resistance, fault distance, fault
type and so on, a number of results of fault location are gained
using assembled algorithm proposed. Results of simulation are
that located errors are within 100m if fault location based on
traveling wave algorithm is available. So it is improved that
assembled algorithm proposed is correct.

Fig.4 Wavelet analysis

IV. CONCLUSION:
Considering roundly reliability and accuracy of
oneterminal fault location system for transmission lines, an
impedance-traveling wave assembled algorithm, which
combines measurement impedance method with traveling
wave method, is presented. The assemble algorithm have
complementarities, because measurement impedance method
guarantees reliability, especially when fault point is very near
to the detective bus, and traveling wave method improves
accuracy, especially when fault point is far from the detective
bus. Results of simulation test have improved that assembled
algorithm proposed is correct. In the case of that exact time of
reflected wave from fault point to detective bus and that of
reflected wave from opposite bus to detective bus can both be
detected, traveling wave velocity has no relation with fault
accuracy, and can be corrected to its actual velocity online. In
the worst situation, measurement impedance method can still
guarantees reliability in despite of that traveling wave method
8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
Fig.3 Model for simulation
285 is no available. So it is prospected that the assemble algorithm
will have a positive effect on fault location technology for

transmission lines of electric power systems and be applied

practically.

V. REFERENCES:
[1] A. Elhaffar and M. Lehtonen: An improved GPS current
traveling-wave fault locator in EHV transmission networks
using few recordings. Presented at Future Power Systems,
2005 International Conference
[2] M. Aurangzeb, P.A. Crossley, P. Gale, "Fault location
using the high frequency traveling waves measured at a single
location on a transmission line," in Proc. 2001 IEE
Developments in Power System Protection, Seventh
International Conf., pp: 403-406.
[3] Darren Spoor and J. G. Zhu, "Improved Single-Ended
Traveling-Wave Fault Location Algorithm Based on
Experience with Conventional Substation Transducers," IEEE
Trans. Power Delivery, vol. 21, pp. 1714-1720, Jul. 2006,
[4] Z. Chen, X. Z. Dong and C. M. Lou, "Robustness of oneterminal fault location algorithm based on power frequency
quantities," in Proc. 2002IEEE, Power Engineering Society
Summer Meeting, vol.3 pp. 1118-1122.
[5] K. Zimmerman, D. Costello, "Impedance-based fault
location experience," in Proc. 2006 IEEE, Rural Electric
Power Conf., pp.1-16
[6] D.W.P. Thomas, C. Christopoulos, Y. Tang, P. Gale and J.
Stokoe, "Single ended travelling wave fault location scheme
based on wavelet analysis," in Proc. 2004 IEE, Developments
in Power System Protection, Eighth IEE International Conf.,
pp. 196 199.
[7] X. Z. Dong, Z. Chen, X. He, K. Wang, and C. Luo,
"Optimizing solution of fault location," in Proc.2002 IEEE
Power Eng. Soc. Summer Meeting, pp. 11131117.
[8] V. Pathirana, P. McLaren, and E. Dirks, "Investigation of a
hybrid travelling wave/impedance relay principle," in Proc.
IEEE Power Eng. Society Transmission and Distribution
Conf., pp. 48-53.
[9] Sant and Y. Paithankar, "Online digital fault locator for
overhead transmission line," in Proc. 1979 IEE, vol. 126,
pp.1181-1185, Nov. 1979.
[10] T. Jiang, Y. P. Lu. "Study of fault locating based on single
traveling waves avoiding wave speed influence," Electric
Power Automation Equipment, vol. 24, pp. 29-32, Dec. 2004.
[11] A. Gopalakrishnan, M. Kezunovic, S.M. McKenna and
D.M. Hamai, "Fault location using the distributed parameter
transmission line model," IEEE Trans. Power Del., vol. 15,
pp. 1169-1174, Oct. 2000.
[12] A. Abur and F. Magnago, "Fault location using wavelets,"
IEEE Trans. Power Del., vol. 13, pp. 1475--1480, Oct. 1988.
[13] G. Kim, H. Kim, and J. Choi, Wavelet transform based
power transmission line fault location using GPS for accurate
time synchronization, in Proc. IEEE Power Eng. Soc.
Transm. And Distribution Conf., vol. 1, 2001, pp. 495-499.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
286

Market behaviour to promote renewable energy

Nuno Domingues
ISEL, Instituto Superior de Engenharia de Lisboa, Portugal
Rua Conselheiro Emdio Navarro, 1, 1959 007 Lisboa; ndomingues@[Link]

Abstract- This paper presents results of a study concerning the

generation strategic bids for a single hour.
In this study I incorporated the price and quantity bids.
I considered an elastic demand curve, approximated by an affine
function, assuming that there is consumers reaction and that the
market price and the demand are related. Also, I consider the
competitors reaction using a parameter that represents the
conjectural variation.
I studied the market behaviour assuming that the market price is
represented by a normal probability function.
I studied and compared the market behaviour for two price
markets types, the MCP (Market Clearing Pay) and PAB (Pay As
Bid), in two situations: without incorporating the externalities and
taking account with the emissions.
Index Terms Strategic Bidding, Generation Surplus,
Conjectural Variation, Elastic Demand, Normal Price Distribution,
Emissions.

I. NOMENCLATURE
*
- block i surplus: mi (.) = mi (ai , ai* , , sell
i , Pgi )

The study of the market behaviour with the conjectural

parameter, developed in 1924 by Bowley and in 1933 by Frisch,
was used by several authors [3], [4], [5] but only to simulate
oligopoly markets with linear bids and determining just one
strategic bid.
The experience shows us that the normal distribution is the one
that best represents the market prices [6]. When we consider a
normal price distribution, the block surplus function is more
complicated than when we consider an uniform price
distribution. [1].
III. METHODOLOGY
I consider a market with several companies that bid by blocks,
each block is identified by i. The block i surplus depends on
both strategic bids: price and quantity. For each strategic bid it is
assumed that all the companies want to maximize the surplus of
each block separately.
I assume that the demand is elastic, allowing the price to change
with the demand. Also, it is assumed that the market price
depends on the demand, as illustrated in Fig. 1:

- block i production cost: ai

- price strategic bid: a

- quantity strategic bid:

- block i selling price,

*
i

Pgi*

sell
i

- expected price assuming a rigid demand:

- maximum expected price assuming a rigid demand: max
- minimum expected price assuming a rigid demand: min

Pmin

Thereby, the market price can be ruled by the equation:

- expected price assuming an elastic demand:

= e s( Pd Pmin )

- maximum expected price assuming an elastic demand: max

- minimum expected price assuming an elastic demand: min

II.

Pd
Fig. 1: Demand curve.

INTRODUCTION

ts desirable that the electricity market work in a perfect

competition. However, due to the limited number of
generation companies (lack of competitors), due to the high
investment (one of the biggest barriers to new players), due to
the long period of time taking from the planning to the
exploration of a plant, the grid capacity and the transmission
losses, the markets tend to work as an Oligopoly. Thereby, some
companies can have a significant market share and make
strategic bids to improve their profit.

(1)

The value e is the maximum price when the demand is equal to

the minimum quantity,

Pmin , and isnt equal to max .

The demand is:

Pd ( Pgi , P- gi ) = Pgi + P- gi ( Pgi )

Where

(2)

P- gi is the aggregated opponents quantity strategic bid.

The value s is the slope of the demand curve and is associated

with the consumers reaction.

According to equation (1) and (2):

= e Karpacz,
s( Pgi +Poland,
P- gi May
Pmin10-13,
) 2009
8. EEEIC International Conference on Environment and Electrical Engineering,
287

Fig. 3: Market Price reliable range according to Pgi .

thereby,

dP-gi
d
= s (1 + ) with =
dPgi
dPgi

The market price with the conjectural variation approach is

(3)

is a parameter which represents the conjectural variation.

This parameter introduces the competitors reaction to the block i

quantity strategic bid. When the block i changes is quantity bid
the competitors change their quantity bid by dP- gi .
Its assumed that

dPgi

P-gi

(4)

Pgi

= s(1 + ) = K s (1 + ) Pgi (5)

Pgi

K is the expected market price for the minimum value of Pgi .

For different values of the parameter

I have:

n = K n s (1 + n ) Pgi
which, assuming

difference to the market assuming a rigid demand.

= + s(1 + )Pgi
Its defined

is constant for each case study. Thereby,

dP-gi

=
then

= + , where = f ( Pgi , ) is the market price

(6)

s constant, can be illustrated by Fig. 2:

n

<1

(8)

Pgi* as

Pgi* = Pgimax - Pgi*

(9)

max = max + s (1 + )Pgi*

*
min = min + s (1 + )Pgi

(10)

Thereby

The quantity strategic bid is

Pgi* [ 0, Pgimax ]. Otherwise, I

consider that the quantity strategic bid is the respective active

restriction.
The selling price,

sell
i , depends on the quantities. In the MCP

market, the active participants payment is equal to the marginal

price. In the PAB market, the active participants payment is
equal to their bid.
For

> 1 ,

when the block reduces is quantity to

Pgi* the

market reacts rising the marginal price to = + . In the

MCP market, the block i will sell less quantity at a higher price.
In the PAB market, the probability of dispatch of the block i is

=1

higher for the blocks that ai > min . For both markets, there
MCP

>1

are a dispatch probability for the block i that ai > max .

MCP

P gi
Fig. 2: Market price variation according to Pgi and

d / dPgi represents the influence of the quantity bid in the

market price, according with
all prices. Thereby,

. Its assumed that it is valid for

d min d max
=
= s (1 + )
dPgi
dPgi
For ,

[ min ,

max

(7)

Thereby, the determination of the strategic generation quantity

and price bids leads to interesting dynamic market behaviour.
Also, since 1 , the demand will change and
Pd = f ( , Pgi ) .
I studied the market behaviour assuming that the market price is
represented by a normal probability function. With the normal
probability function I assumed that the market price has higher
probability to be in the middle of the reliable range
[ min , max ], as shown by Fig. 4

] and can be illustrated by Fig. 3.

f ( )

max

min

max

Fig. 4: Truncated normal probability price function

min

The normal function has an error when its limited by the range

Pgi

[ min , max ] since its defined for all domain. However, I

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
288

assume that 8 = max - min , thereby the maximum error will

be 0,006% [6].

I consider that the strategic price bid of the block i doesnt

influence the market price, thereby

The price probability distribution is then

f ( ) =

1 2
)
(
2

; f ( ) ~ N (

max + 3min max min

2

Thereby we have

2e J x

( max min )

In the MCP market, the selling price,

8(2 + max + min ) 2

Jx =
(max min ) 2
Its also defined an auxiliary variable

8(2ai* + max + min ) 2

( max min ) 2

f ( )(u( - min ) - u( - max ))(isell - ai )(Pgimax - dPgi )d

where u(.) is the step function.

For the block i, the strategic bids are determined by the
resolution of the following maximization problem:
*
max mi (a i* , a i , , sell
i , Pgi )

a i* , Pgi*
s.t.
*
gi

max
gi

P -P
min
gi

*
gi

-P 0

a i - a i* 0
I assume only the production limits restriction for the quantity
strategic bid.
To avoid negative surplus for the block i, I assume that the price
If the surplus function is concave and the restrictions are not
active, the strategic bids can be determined by:

a * mi (.) = 0
i

mi (.) = 0
Pgi*

PgiMCP . Based on the

production cost, the strategic bids are

MCP

1) if ai > max

aiMCP = a i
.
MCP
max
Pgi 0, Pgi

2) if

MCP
MCP
min
< ai max

First option
MCP
max
MCP MCP
max + min + 3ai + 2 s (1 + ) Pgi
=
ai
5

MCP
MCP

+
2
PgiMCP = Pgimax + max min 2ai

5
5s(1 + )

Or second option

aiMCP = ai

MCP
max
ai
MCP
max
P
P
=
+
gi
gi
s(1 + )

According to the option that leads to higher expected surplus for

block i.

strategic bid is always a i a i .

, is the marginal price,

. The quantity strategic bid that

maximizes the expected block i surplus is

According to the production cost, the block surplus function is:

+

sell
i

MCP . The price strategic bid a i* that maximize the expected

MCP

= 0.

IV. CASE STUDY

block i surplus is a i

where

mi (.) =

ai *

1) For the MCP market

f ( ) = 4

J =

3) if a i

MCP
min

MCP
ai aiMCP min

max
MCP
MCP
MCP
MCP Pgi
min + max 2a i
=
+
P
gi

2
4s (1 + )

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
289

According to the strategic bids, the maximum expected surplus

for block i is
a) if a

MCP
i

MCP
max

>

According to the production cost, the strategic bids are

PAB

1) if ai > max

aiPAB ai
PAB
max
Pgi 0, Pgi

mi MCP max (.) = 0

MCP
MCP
min
< aiMCP max

b) if

mi MCP max (.) =

8 2

MCP
((e 8 - e 8 ( Jaux ) )( MCP
max - min ) +

2) if
8+8( Jaux ) 2

+ (e

)(

MCP
max

ai 2 max
+ Pgi ) +
s(1 + )
5

+ 2e 8+8( Jaux)

PAB
PAB
min
< ai max

First option
PAB
PAB
PAB 3max
+ 3min
+ 2ai + 2 s (1 + )PgiPAB
=
a
i

PAB
PAB
+
+
2
(
1 + ) Pgimax 2aiPAB

s
max
min
P PAB = 3
gi
8s (1 + )

2 ( MCP
max +

16 2 Jaux
+ MCP
+
a
ERF
ERF
2
)(
(
2
2
)
)))
( MCP
i
min
MCP
max min

Or second option
PAB
PAB
PAB max
+ min
+ 3ai + 2 s(1 + ) Pgimax
a
=
i
5

PAB 2 max PAB + PAB 2a

max
i
Pgi = Pgi + max
5
5s(1 + )

where for the first condition

Jaux =

MCP
max
3( MCP
max + min ) 6 a i + 4 s (1 + ) Pgi
MCP
5( MCP
max min )

According to the option that leads to higher expected surplus for

block i.

and for the second condition

Jaux =

MCP
max
+ MCP
min 2a i
,
MCP
max MCP
min

PAB

3) if ai min

ERF ( X ) is the integral of the Gaussian distribution, given by

ERF ( X ) =

c) ai

MCP

(1) n X 2 n +1
.

n =0 (2n + 1)n!

t
e dt =
2

mi MCP max (.) = Pgimax (

MCP
MCP
max + min
2

ai )( ERF (2 2 ))

In the MCP market, the selling price,

PAB

If

sell
i

, is the price strategic

. The quantity strategic bid that maximizes the

expected block i surplus is

PAB
gi

PAB
ai < yPAB then aiPAB = min
.

PAB

I define

< ai

Y =

PAB
max

then

PAB
i

PAB
PAB
3 max
+ 3 min
+ 2ai
=
8

5,5 min - 3max

as an auxiliary variable.
2

According to the strategic bids, the maximum expected surplus

for block i is

2) For the PAB market

PgiPAB = Pgimax

If

MCP
min

bid, ai = a i

a) if ai

PAB

PAB
> max

m PAB max (.) = 0

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
290

b)

PAB
PAB
min
< aiPAB max

mi

PAB max

1
PAB
max
(.) =
( PAB
max + min + 2 s (1 + ) Pgi
250s (1 + )

PAB
PAB
2ai )(3( max
+ min
) + 16s (1 + ) Pgimax )( ERF (2 2 ) -

ERF (

c)

PAB
PAB
2 2 (9(max
- min
) - 18ai - 2s (1 + ) Pgimax
PAB
PAB
25(max
- min
)

))

VI. CONCLUSIONS
I assume that the companies have price and quantity strategic
bids to maximize their surplus. According to the tables in
appendix, we can see that the influence of all technologies is
bigger in the MCP market than in the PAB market, when I
assume a normal price distribution. The demand satisfied is
lower and the market price is higher in the MCP market than in
the PAB market.
Also, when the emission externality is introduced as a
production cost, the surplus is lower. Therefore, the market can
work as an incentive for sustainability.

PAB
YPAB < a iPAB min

mi PAB max (.) =

d) ai

PAB

mi

max
gi

REFERENCES

PAB
( PAB
max + 3 min

6a i )( ERF ( 2 2 ))

[2]

YPAB

PABma

[1]

[3]

(.) = Pgimax ( PAB

min

ai )( ERF (2 2 ))

V. RESULTS

[4]
[5]

The results are for the following cases:

[6]

Domingues, N., Mendes, V., Correia, P., Oferta Estratgica da Gerao em

Mercados de Energia, ISEL, 18 e 19 de Outubro de 2005, 1st Workshop in
ISEL;
Ren, Y., Galiana, F., Pay-as-bid versus Marginal Pricing- Part I: Strategic
Generator Bid, IEEE Transactions On Power Systems, Vol. 19, No. 4,
November 2004;
Song, Y., Hou, Z., Wen, F., Yixin, N., Flix, F., Conjectural Variation
Based Learning of Generators Behaviour in Electricity Market, IEEE
Transactions On Power Systems, Vol., No. , 2003;
Day, C., Hobbs, B., Pang, J., Olipolistic Competition in Power Networks:
a Conjectural Supply Function Approach, IEEE Transactions on Power
Systems, Vol. 17, n3, Ago 2002;
Song, Y., Ni, Y., Wen, F., Wu, F., Analysis of strategic Interactions
Among Generation Companies Using Conjectured Supply Function
Equilibrium Model, IEEE, 2003;
Montgomery, D., Runger, G., Applied Statistics and Probability for
Engineers, Wiley, 2007.
[Link]

0.0001

-0.9900

1.0000

0.0000

VII. BIOGRAPHY

1.0000

0.5000

1.0000

1.0000

1.0000

1.5000

1.0000

2.0000

Nuno A. S. Domingues (b. 1972) received the Licenciate (5-year) degree in

Electrical Engineering from ISEL (2005) and Master degrees in Electrical
Engineering and Computer Science from IST (2008). Presently he is a PhD
candidate at FCT with his thesis on energy, sustainability and efficiency.
Presently he works as a Labour Engineer at ISEL (2001-2008) in
Engineering Systems.
His topics of research include electricity markets modelling and simulation,
energy systems, sustainability, efficiency.

Case

[7]

Fig. 5: Case Study.

The emissions of a coal power plant are 1000kg/MWh [7],

thereby the cost of introducing the emission externality is 20
/MWh.
The results were obtained for the following values:
- for a i = 15 , coal technology without taking account with the
externalities;
- for ai = 35 , coal technology taking account with the
externalities;
-

min = 22 ; max = 38 .

The results are in the

Appendix.
8. EEEIC
International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
291

I. APPENDIX

a i = 15

MCP Market
Bid

Surplus

Case

Price

Quantity

15,00

10,00

15,00

15,00
15,00

8,75

16,25

7,50

18,75

15,00
15,00

6,88

5
6

15,00

3
4

15,00

Per unit

PAB Market
Market Price

Total

Bid

Surplus

Market Price

Minimum

Maximum

Price

Quantity

Per unit

Total

Minimum

Maximum

149,99

22,00

38,00

28,13

10,00

10,84

108,38

22,00

38,00

142,18

23,25

39,25

28,13

10,00

10,84

108,38

22,00

38,00

140,62

25,75

41,75

28,13

10,00

10,84

108,38

22,00

38,00

21,25

146,08

28,25

44,25

28,13

10,00

10,84

108,38

22,00

38,00

6,50

23,75

154,37

30,75

46,75

28,13

10,00

10,84

108,38

22,00

38,00

6,25

26,25

164,05

33,25

49,25

28,13

10,00

10,84

108,38

22,00

38,00

5,63

41,25

232,02

48,25

64,25

28,13

10,00

10,84

108,38

22,00

38,00

Table. 1: Market behaviour without externalities

ai = 35

MCP Market
Bid

Surplus

PAB Market
Market Price

Bid

Surplus

Market Price

Case

Price

Quantity

Per unit

Total

Minimum

Maximum

Price

Quantity

Per unit

Total

Minimum

Maximum

35,00

10,00

0,004

0,04

22,00

38,00

35,25

10,00

0,002

0,02

22,00

38,00

37,00

2,00

2,78

5,56

30,00

46,00

35,50

4,69

0,04

0,19

27,31

43,31

37,00

2,67

6,00

16,00

33,00

49,00

35,50

5,00

0,11

0,55

29,50

45,50

35,00

2,25

10,50

23,62

37,50

53,50

36,25

5,16

1,20

6,17

31,69

47,69

35,00

2,40

14,00

33,60

41,00

57,00

37,50

5,25

2,46

12,94

33,88

49,88

35,00

2,50

17,50

43,75

44,50

60,50

38,50

5,31

3,74

19,84

36,06

52,06

Table. 2: Market behaviour with emission externalities

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
292

The future goes offshore

Robert Hoyer, Henry Rmer, and Jonathan Saudhof

AbstractWithin the renewable energies wind parks are

playing an important role. Since onshore wind energy
market is already high developed and concerns are more
about repowering, offshore wind plants have gained
enormous popularity. In consideration of the high
potential German government set the target to
accommodate 15 percent of the energy demand with the
help of offshore wind plants in 2030. Despite the high
potential there are still risks and problems hindering the
installation of huge offshore plants. In fact there is still no
such wind park completely installed in Germany.

the installed number of turbines in Germany is irrelevant to

the United Kingdom with 591 MW, Denmark 426,35,
Netherlands with 246,8 MW and Sweden with 132,5 MW [2].
The prognosis says that a possible capacity of 20000 up to
30000 MW can be realized. In long terms offshore wind
power will affect the market of renewable energy sources [4].

Index TermsEconomic potential, offshore, Renewable

energy, wind energy
I. INTRODUCTION

Germany is one of the leading countries in using renewable

energy. The German government and others try to reduce their

CO emissions rapidly. Besides expanding solar- and
photovoltaic techniques the main energy source is wind
power. 2005 4.3 % of the produced energy were generated by
onshore windmills [3]. In the future the expansion and new
installation of those wind farms is limited due to the present
legal conditions for ecological safety. In this context new
ways are going to be approached. Offshore Wind parks are
actually the best choice to extend the use of wind power
simultaneously with an increase of efficiency.
The advantages of these offshore solutions in contrast to wind
parks are the stable and higher level of wind speed. Currently
the using of wind power is connected with occurring gaps
between generated power and the actual system load due to
the onshore wind speed fluctuation.
Actually in Germany no offshore projects have been realized
yet. 2008, the installation of the first test park alpha ventus
has been started. Twelve wind turbines with 5-6 MW turbines
are going to be built in 30m depth [1]. In the international
comparison of realized offshore projects Germany shows that

Fig. 1 Developement of renewable energy

Chances:
Wind speed and development:
As mentioned before the less fluctuation and higher wind
speed is a major advantage of offshore wind parks. This
means a push of generating efficiency and the possibility that
wind turbines can supply with base load current.
Figure 2 shows the differences in wind speed per minute
depending on the geographical location and gives an inside of
benefits from those installations.

Manuscript received March 20, 2009.

Robert Hoyer is a student of industrial engineering at the
Brandenburgische
Technische
Universitt
Cottbus,
Germany
(e-mail: [Link]@ [Link]).
Henry Rmer is a student of industrial engineering at the Brandenburgische
Technische
Universitt
Cottbus,
Germany
(e-mail: [Link]@ [Link]).
Jonathan Saudhof is a student of industrial engineering at the
Fig. 2 Offshore wind conditions in Europe
Brandenburgische
Technische
Universitt
Cottbus,
Germany
(e-mail: [Link]@ [Link]).
8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
293

2
The future of windmills in Germany is in offshore farms. New
installations of onshore turbines will be stopped due to the
present legal conditions for ecological safety. This means a
limited size for turbines because of their air noise emission
and the casting of shadow. The target for Germany is to
explore and lead the market after the United Kingdom within
the next years. Figure shows the size of installed turbines
planed till 2015.
Parallel to reducing new installations repowering is going to
be the key for increasing the efficiency of existing windmills.
Learning effects in various technical sciences will determine
the development of future turbines. Material science,
Aerospace and experiences in the installation of offshore
platforms will highly influence the future offshore wind parks.
A project actually presented in the press called FloDesign
Wind Turbine is a new way for designing windmills to
maximize the used wind power.

for gas and steam power stations established in 2005 at 3.6

cent per kWh. For brown coal-fired power stations the IER
comes on approximately 3.1 and for coal-fired power stations
on approximately 2.8 cent per kWh. According to the
external investigation done by the IER in Germany, the
external costs of the stream production of coal or brown coal
amount from 3 to 6 cent per kWh. Wind energy onshore or
offshore - has extremely low external production costs in
comparison.
Average stream actual costs
in the EU

4 of cent/kWh

In 2005 established gas and

steam power stations

3,6 of cent/kWh

Brown coal-fired power

stations

3,1 of cent/kWh

Coal-fired power stations

2,8 of cent/kWh

Additional external costs for

coal-fired power stations and
brown coal-fired power
stations in Germany

3 6 of cent/kWh

b)

Fig. 3 Operating offshore wind farms

Present Offshore experiences

For the first smaller Offshore-wind parks close to coast stream

actual costs have been determined abroad in the area from 6 to
8 cent. Besides, the later projects show a sinking trend. The
DEWI (German Wind Energy Institute) has published
estimates, for which the planned wind parks in Germany,
standing 40 kilometers far from the coast remotely will reach
stream actual costs of from 7.4 to 8.1 cent per kWh. At the
moment installed power is 2 MW [5]. However, the
applications planned from 4 to 5 MW arrangements what will
lead probably to a lowering of the costs in the future.

II. ECONOMY
The initial costs of wind energy onshore have drastically
decreased because of the gradual technical advancement of the
arrangements. In Germany the costs of wind energy
arrangements per kilowatt of achievement were lower
according to the federal association Windenergie inc. in 2000
almost
around
half
than
still
in
1990.
According to the European wind energy association EWEA
the actual costs of modern wind strength arrangements in
good wind locations are at the moment to 5.1 cent / kWh. Up
to 2013 3.2 of cent / kWh are accessible according to this
association in such locations [5].
a)

Actual cost of power generation

In 2001 the external research project of the European

commission came to the result that the average stream actual
costs amount to 4 of cent / kWh in the European Union.
The Institute of energy industry and rational energy use (IER)
of the University of Stuttgart estimates the stream actual costs

Fig. 4 Offshore wind markets in Europe

III. RISKS
Despite the fact offshore wind energy to be a solution with
high potential within the renewable energies it is still facing a
variety of problems to solve. On the one hand there several

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
294

3
concerns about how offshore wind parks may influence the the peril of collisions with oil tanker leading to environmental
maritime environment negatively. Among conservationists the catastrophes.
threat of collision of birds with rotor blades are mentioned as
IV. CONCLUSION
well as their displacement from their native environment. In
fact there is still no scientific evidence of a high rate of
Although Germany is supposed to be the vanguard
collisions. Indeed Danish researchers with the help of their considering renewable energies, especially in wind energy it
Thermal Animal Detection System (TADS) could show that has lost this role in the field of offshore wind parks just by
the number of incidents is quiet small. Nevertheless this taking a look at the number of installed plants. First
problem is considered to be locally different. But not only commercially used plants are already running near the costs of
birds seem to be threatened. Sea fish and marine mammals its European neighbors Great Britain and Denmark [10]. At
may suffer the artificially created magnetic and electric fields least there is no doubt offshore wind energy will face a high
by the cables connecting the plants offshore with the electrical growth during the next years in all countries verging on to the
network. For this reason these animals such as whales might sea. Nevertheless with the construction of the Alpha Ventus
lose orientation. Additionally benthos are losing their habitats wind park by the German RWE a large step for the future of
because of the building of the fundament at the ground of the offshore wind energy has been done.
sea [6]. Nevertheless research in this field is ongoing, e.g.
considering the research station Fino I near the Island of
ACKNOWLEDGMENT
Borkum, Germany and there is still no conclusion [8]. But
Robert Hoyer would like to thank his parents, grandparents
there are not only environmental concerns to deal with. The
and God. He specially thanks Rommy for the loving support
installation of wind parks is still not only a technical challenge
all the time. Henry Rmer thanks his family especially his
but a financial risk as well [7]. A wind park planed consisting
brother Stephan for reading proof of this paper. Jonathan
of 150 to 180 plants by the German Enterprise RWE near the
Saudhof would to thank all of his friends and his family.
island of Juist will take an investment of ca. 2.8 Billion Euro.
Aware of that fact high investments are needed only the huge
REFERENCES
energy provider will be able to build the first commercially
used offshore wind parks. The installation of wind parks is [1] Dena (2009, February). Internationale Windprojekte:
[Link]
much more complicated and referring to that expensive.
[2] Deutsche Windindustrie in Deutschland (2009, February).
Installing a 1 Megawatt plant offshore takes about 2.5 Million
Onshore-Windenergie und Repowering: [Link] two times the price for a plant to build onshore [9].
[Link]/fakten/onshore/[Link]
According to a study by Price Waterhouse Coopers (PWC) [3] Dena (2009, Februar). Offshore-Testfeld alpha ventus:
financing offshore projects suffer two significant problems.
Deutschlands erster Windpark auf dem Meer:
[Link] there has been a high increase of investing costs due to
[Link]/page/[Link]?id=9352&L=1&fs=1&L=0
higher price for natural resources, especially steel and copper.
[4] Kller, Julia. Offshore wind energy. Berlin: Springer,
Furthermore the increasing demand for onshore plants has
2006
raised the price for offshore plants too since these are
[5] Dena. (2009, February). Fakten Wirtschaftlichkeit.
technically based on onshore technologies although there have
Available:[Link] built prototypes specially designed for offshore
[Link]/page/[Link]?id=2601
operation. Additionally the uncertainty of investment and [6] T. Merck, H. v. Nordheim (1999) Probleme bei der
maintenance cost at the beginning of a project is still an
Nutzung von Offshore-Windenergie aus Sicht des
obstacle for potential investors considering that the
Naturschutzes. German Journal of Hydrography
experiences regarding wind parks far from the shore in thirty
Supplement 10
meter depth are low. The second issue concerns the so called [7] J. Ondraczek, H. Stohlmeyer, S. Kver (2006).
interface problem. An offshore project joins a number of
Finanzierung der Offshore-Windenergie in Deutschland:
enterprises for building the wind park whereas onshore plants
Probleme
und
Lsungsanstze.
Available:
are built by one corporation normally. Next to the plant the
[Link]
building of the electrical and structural connection has to be [8] D. Asendorpf. (2008, June). Viel Energie auf hoher See.
Available: [Link]
done too. But these different works refer to different amount
of risk. The organization of the project must guarantee a fair [9] M. Uken (2006, November). Riskante Wellenbrecher.
Available: [Link] of risk for any enterprise within it in order to
Windkraft
make enterprises invest and reduce risk surcharge. A very
important point determining the economic success of offshore [10] RWE (June, 2008). Fact Book Renewable Energy
Available:[Link]
projects refers to maintenance. Because plants are located at
ovationen/services/infothek-forschung-entwicklung/
difficult accessible locations availableness is threatened in the
case of technical breakdown. So, in fact there might be the
problem of delayed repairs when weather conditions are bad
8. EEEIC
International
Conference of
on Environment
[7]. Another risk
pronounced
by inhabitants
cost regionsand
is Electrical Engineering, Karpacz, Poland, May 10-13, 2009
295

Strong Electromagnetic Impulse Generation

Daniel Pyda

Wroclaw University of Technology

27 Wybrzeze Wyspianskiego
Wroclaw, Poland

Abstract - Electromagnetic fields can have not only disturbing

but also damaging influence on equipment. There are many ways
to generate such electromagnetic field. In this paper some of
electromagnetic field generators will be introduced and how
dangerous such field is for today electric and electronic equipment.

Earth's magnetic field at high latitudes. These new

calculations, combined with the accelerating reliance on EMPsensitive microelectronics, heightened awareness that the EMP
threat could be a very significant problem.
In 1962, the Soviet Union also performed a series of three
EMP-producing
nuclear tests in space over Kazakhstan called
I. INTRODUCTION
"the k project". Although these weapons were much smaller
The fact that an electromagnetic pulse is produced by (300 kilotons) than the starfish prime test, since those tests
a nuclear explosion was known since the earliest days of were done over a populated large land mass (and also at a
nuclear weapons testing, but the magnitude of the EMP and the location where the earth's magnetic field was greater), the
significance of its effects were realized very slowly.
damage caused by the resulting EMP was reportedly much
In July, 1962, a 1.44 megaton united states nuclear test in space, greater than in the starfish prime nuclear test. The geomagnetic
400 km. Above the mid-pacific ocean, called the starfish prime storm-like E3 pulse even induced an electrical current surge in
test demonstrated to nuclear scientists that the magnitude and a long underground power line that caused a fire in the power
effects of a high altitude nuclear explosion were much larger plant in the city of Karagandy. After the collapse of the soviet
than had been previously calculated. Starfish prime also made union, the level of this damage was communicated informally
those effects known to the public by causing electrical damage to scientists in the united states. Formal documentation of some
in Hawaii, more than 800 miles away from the detonation point, of the EMP damage in Kazakhstan exists but is still sparse in
knocking out about 300 streetlights, setting off numerous the open scientific literature.
burglar alarms and damaging a telephone company microwave
II. CHARACTERISTIC
link.
The EMP damage of the starfish prime test was
The case of a nuclear electromagnetic pulse differs from
quickly repaired because of the ruggedness (compared to today) other kinds of electromagnetic pulse (EMP) in being a complex
of the electrical and electronic infrastructure of Hawaii in 1962. electromagnetic multi-pulse. The complex multi-pulse is
Realization of the potential impacts of EMP became more usually described in terms of 3 components, and these 3
apparent to some scientists and engineers during the 1970s as components have been defined as such by the international
more sensitive solid-state electronics began to come into standards commission called the International Electrotechnical
widespread use.
Commission (IEC).
The larger scientific community became aware of the
The 3 components of nuclear EMP, as defined by the IEC,
significance of the EMP problem after a series of three articles are called E1, E2 and E3.
was published about nuclear electromagnetic pulse in 1981 by
The E1 pulse is a very fast pulse that generates very high
William broad in the weekly publication science.
voltages. E1 is the component that can destroy computers and
The relatively small magnitude of the starfish prime EMP in communications equipment and is too fast for ordinary
Hawaii (about 5,700 volts/meter) and the relatively small lightning protectors.
amount of damage done (for example, only 1 to 3 percent of
The E2 component of the pulse many similarities to the
streetlights extinguished) led some scientists to believe, in the electromagnetic pulses produced by lightning. Because of the
early days of EMP research, that the problem might not be as similarities to lightning-caused pulses and the widespread use
significant as was later realized. Newer calculations showed of lightning protection technology, the E2 pulse is generally
that if the starfish prime warhead had been detonated over the considered to be the easiest to protect against.
northern continental united states, the magnitude of the EMP
The E3 component of the pulse is a very slow pulse,
would have been much larger (22,000 to 30,000 volts/meter) lasting tens to hundreds of seconds, that is caused by the
because of the greater strength of the earth's magnetic field nuclear detonation heaving the earth's magnetic field out of the
over the united states, as well as the different orientation of the way, followed by the restoration of the magnetic field to its

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
296

natural place. The E3 component has similarities to a

geomagnetic storm caused by a very severe solar flare.
III. EXPLOSION ON EARTH SURFACE
During the case of a nuclear explosion on Earth surface
(Fig.1) the symmetry conditions arent preserved. We can treat
earth as a perfect absorber for radiation and current conductor.
Because of that un symmetry a current vertical component is
forming which generates an electromagnetic field like vertical
monopole. Electromagnetic energy is been radiated beyond
Figure 2. Earth nuclear explosion electromagnetic impulse shape
deposition region.
The conditions of electromagnetic impulse propagation are
in accordance with theory of wave propagation near Earth.
Some of electrons that propagates radial from place of
explosion returns to this place through ionized magnetic paths
in air and Earth. Therefore current circles are forming and
produce strong azimuth magnetic field. The electric field made
during the explosion is very high but it decreasing quite fast, at
first inversely proportional to power third of distance. In bigger
distances from influence area we engage to so called far zone
and field decrease inversely proportional to distance from
explosion place. The range of influence for electromagnetic
field generated in such conditions is relatively small reaching
Figure 3. The in atmosphere explosion
several kilometers. Typical on Earth nuclear explosion
electromagnetic impulse shape is shown on (Fig.2).
Main parameters of electromagnetic impulse can be shown
Main parameters of electromagnetic impulse can be shown as below:
as below:
- impulse increasing time < 50s
- impulse increasing time < 5ms
- impulse lasting time < 100s
- impulse lasting time < 200ns
- E < 10kV/m
- E < 40kV/m
- H < 50A/m
- H < 200A/m
IV. EXPLOSION IN ATMOSPHERE
During the case of a nuclear explosion in atmosphere (Fig.3)
equivalent electric dipole is been formulating, what is caused
by atmosphere inhomogeneity, which is a radiation producer.
Some of this radiation is been reflected from Earth surface and
interfere with direct radiation explosion electric field has lower
intensity than the on Earth explosion field.

Figure 1. Nuclear explosion on Earth surface

V. EXPLOSION IN GREAT ALTITUDE

During the case of a nuclear explosion in great altitude (Fig.4),
recoil electrons are produced only when photons reaches the
thicker parts of atmosphere. These electrons are deflect by
Earth magnetic field producing crosswise directed electric
current, which is the source of electromagnetic radiation
propagating radial from hot spot. Such rising electromagnetic
impulse (HEMP High Altitude EMP) has great range (Fig.5).

Figure 4. Nuclear explosion in great altitude

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
297

Figure 7. USS Estocin (FFG-15) moored near the Electro Magnetic

Pulse Radiation Environmental Simulator for Ships I (EMPRESS I) facility.
(Antennae at top of image)
Figure 5. Nuclear explosion in great altitude damage range

Main parameters of electromagnetic impulse can be shown

as below:
- impulse increasing time < 10ms
- impulse lasting time < 100ns
- E < 100kV/m
- H < 300A/m
Parameters and maintenance of electromagnetic produced
during nuclear explosion in great altitude are base for
determine parameters and time run for conventional
disturbance ( test signal ), recommended by many standards to
perform tests for equipment sensitivity for electromagnetic
field.
VI. NON-NUCLEAR ELECTROMAGNETIC IMPULSE
Non-nuclear electromagnetic pulse (NNEMP) (Fig.6-7) is
an electromagnetic pulse generated without use of nuclear
weapons. There are a number of devices that can achieve this
objective, ranging from a large low-inductance capacitor bank
discharged into a single-loop antenna or a microwave generator
to an explosively pumped flux compression generator. To
achieve the frequency characteristics of the pulse needed for
optimal coupling into the target, wave-shaping circuits and/or
microwave generators are added between the pulse source and
the antenna. A vacuum tube particularly suitable for
microwave conversion of high energy pulses is the vircator.

Figure 6. A right front view of a Boeing E-4 advanced airborne command

post (AABNCP) on the electromagnetic pulse (EMP) simulator for testing.

NNEMP generators can be carried as a payload of bombs and

cruise missiles, allowing construction of electromagnetic
bombs with diminished mechanical, thermal and ionizing
radiation effects and without the political consequences of
deploying nuclear weapons.
NNEMP generators also include large structures built to
generate EMP for testing of electronics to determine how well
it survives EMP. In addition, the use of ultra-wideband radars
can generate EMP in areas immediately adjacent to the radar;
this phenomenon is only partly understood.
Information about the EMP simulators used by the United
States during the later part of the Cold War is now in papers
under the care of the SUMMA Foundation, which is now
hosted at the University of New Mexico.
The SUMMA Foundation web site includes documentation
about the huge wooden Trestle simulator in New Mexico,
which was the world's largest EMP simulator which used a
specialized version of a Marx generator.
VII. EXAMPLE OF VIRCATOR GENERATOR
(Fig.8) A vircator (VIRtual CAthode OscillatOR) is a
microwave generator that is capable of generating brief pulses
of tunable, narrow band microwaves at very high power levels.
A typical vircator is built inside an evacuated resonant cavity
or waveguide. An electrode at one end injects an intense
electron beam, such as from a Marx generator or a flux
compression generator. The electrons are attracted to a thin
anode, such as an aluminized PET film, that is connected to the
grounded waveguide body.

Figure 6. A right front view of a Boeing E-4 advanced airborne command post
(AABNCP) on the electromagnetic pulse (EMP) simulator for testing.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
298

The unit is surrounded by a magnet. Due to the intensity

of the electron beam, many electrons pass through the anode
into the region beyond it, forming a virtual cathode. The
electron beam must be intense enough to exceed the space
charge limiting current in that region, causing oscillations that
generate microwaves. The frequency, efficiency and other
characteristics of the emitted beam depend on the precise
physical configuration and operating parameters.
Vircators have been used as electromagnetic pulse generators
and for generating X-rays. Power levels on the order of 1010 to
1012 watts are possible.
REFERENCES
[1]

Donald J. Sullivan, "High Power Microwave Generation From a Virtual

Cathode scillator (Vircator)," IEEE Trans. Nucl. Sci., vol. NS-30, No. 4,
3426-3428 (Aug. 1983)
[2] [Link]
[3] Tadeusz [Link],Electromagnetic Compatibility Examination
For Electric and Electronic Equipment

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
299

Detection of Fault Position with Respect to

the Compensating Bank in Series Compensated
Line by Measurement of Phase Shift for
Distance Relay Input Currents
Piotr Maniewski, Jan Iykowski
Wroclaw University of Technology (I-8), Wyb. Wyspianskiego 27, 50-370 Wroclaw, Poland
(email: [Link]@[Link], [Link]@[Link])
Abstract--This paper presents an analysis of possibility of
detection of a fault position with respect to the compensating bank
in a series compensating transmission line. The algorithm
designed for this purpose is based on determining the change of
phase shift for the distance relay input currents. Fuzzy logic
technique is applied for making the decision whether a fault is in
front of the compensating bank or behind it. The algorithm has
been tested and evaluated with use of the fault data obtained from
versatile ATP-EMTP simulations of faults in the test network
containing the 400 kV, 300 km transmission line, compensated
with the aid of the compensating bank installed at mid-line. The
sample results of the evaluation are reported and discussed.

I.

INTRODUCTION

Increased transmittable power, improved power system

stability, reduced transmission losses; enhanced voltage control
and flexible power flow control are the reasons behind
installing Series Capacitors (SCs) on long transmission lines
[2]. The environmental concerns stand for that too.
Both, capacitors of fixed value (FSC Fixed Series Capacitors) and of controlled
value (TCSC Thyristor Controlled Series Capacitors) are installed in
series compensated lines. This paper deals with distance protection issues for a line compensated with a
three-phase bank of fixed series capacitors (SCs) installed at mid-line (Fig. 1). SCs are equipped with

their overvoltage protection devices: typically Metal Oxide

Varistors (MOVs). Each MOV is in turn protected from
overheating with the aid of the thermal protection (TP), which
eventually sparks the respective Air-Gap, in order to by-pass
its MOV.
B
A
SYSTEM B
SYSTEM A
FB
FA
SCs
CTsA

MOVs
VTsA

{iA}
{vA}

DISTANCE
RELAY

AIR-GAPs
TP

Figure 1. Schematic diagram of series compensated line for protection study:

FA fault in front of SCs/MOVs, FB fault behind SCs/MOVs.

The compensating bank when installed in a line creates,

however, certain problems for its protective relays and fault
locators. If a series compensated line suffers a fault behind the
SCs, as seen from the relaying point (fault FB in

Fig. 1), a fault loop measured by a distance relay contains,

depending on a type of fault, one (for single phase faults) or
even two (for interphase faults) systems of SCs and MOVs.
As a consequence, the operating conditions for protective
relays become unfavourable and include such phenomena as
voltage and/or current inversion, subharmonic oscillations,
high frequency oscillations due to MOVs [2]. The most
important peculiarity of a series compensated line as the object
to be protected, lays, however, in the fact that the positive
sequence impedance measured by a traditional distance relay is
no longer an indicator of the distance to a fault. The SC and its
MOV affect both, the steady state and transient conditions of
the distance relay measurements.
Protection of networks with series compensated lines is
considered as one of the most difficult tasks. There is still
much room for developing efficient protective relaying for
such networks. The approach presented in this paper is one of
the attempts for realizing that [1], [3], [11].
In this paper a new criterion of qualification of a fault place
with respect to the compensating bank is introduced. It is based
on measuring the change of a phase shift of input currents of a
distance relay.
II. FUZZY LOGIC BASED ALGORITHM FOR DETECTING FAULT
POSITION WITH RESPECT TO COMPENSATING BANK
The concept of fuzzy logic was conceived by Lotfi Zadeh as
a way of processing data by allowing a partial set membership
rather than a crisp set membership or non-membership. It
offers several unique features that make it a particularly good
choice for many problems [6], [9].
A classic logic is based on two values, the most often
represented by 0 and 1, or truth and falsehood. A border
between them is definite unambiguously and invariable. Fuzzy
logic introduces values between the standard 0 and 1 and
fuzzification of the borders between them, giving the
possibility of appearing the values between these values.
Schematic diagram of the proposed fuzzy logic algorithm is
presented in Fig. 2.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
300

u ph (n )

Making the decision on the fault position is a result of the

fuzzy comparison determined with the following relationship:

Ppre>0

Calculation
of power

Ppre<0

i ph (n )

Measurement
of phase shift

(n )

Fuzzification

Setting_1

((n ))

Setting_2

W=

F1
F

where:
F area below the membership function

Fuzzy
Comparison

Decision

Figure 2. Schematic diagram of fuzzy logic based detection of fault position.

(3)

( ) ,

F1 area obtained from F by limiting it with the applied setting

(Fig. 3).
If the ratio (3) exceeds the specified threshold, the decision
is taken that a fault is in front of the compensating bank;
otherwise a fault is classified as behind the bank.

First, digital measurement of phase shift for phase currents is

performed [8]:

III. SIMULATIONS AND TESTING CONDUCTED WITH USE OF

MATLAB AND ATP-EMTP

imag(i ph ( n ))
(n ) = atan

real(i ph ( n ))

The presented algorithm for detecting the position of the

fault with respect to the compensating bank, in terms whether a
fault occurred in front or behind the bank, has been tested and
evaluated
with
the
fault
data
obtained
from
versatile ATP-EMTP [4] simulations of faults in the test
network (Table I).
The test power network contains the 400 kV, 300 km
transmission line, compensated with a three-phase bank of
series capacitors installed at mid-line. The compensation rate
of 70% was assumed. MOVs installed in parallel to series
capacitors were modelled as nonlinear resistors defined with
the analytical characteristic and its parameters as given in
Table I. The thermal protection (TP in Fig. 1) preventing the
MOV from overheating was modelled as the component
integrating the accumulated energy. After exceeding the set
threshold for energy, the sparking of the associated air-gap
undergoes, and the MOV becomes shunted. It has been
checked that for the air-gap sparking does not take place prior
to detecting a fault position. Thus, the thermal protection does
not influence the algorithm.
The model includes the Capacitive Voltage Transformers
(CVTs) and the Current Transformers (CTs). The analogue
filters with 350 Hz cut-off frequency were also included. The
sampling frequency of 1000 Hz was applied.
Faults, occurring in front and behind the compensating bank,
of different specifications have been modelled:
distance to fault 17 values: 30, 40, 45, 50, 60, 90, 120,
149, 151, 180, 210, 240, 250, 255, 260, 270, 290 [km],
fault resistance:
for faults LE and LLE 5 values: 0.1, 5, 10, 25,
50 [],
for faults LL, LLL (LLLE) 5 values: 0.1, 0.5, 1,
2, 5 [],
point on the wave at which the fault is applied 10 cases:
by changing the angle from 0o (at zero crossing) up to 90o
(at maximum),
angle of EMFs at the end A was set as =0 (the cosine
wave with zero phase shift in the phase a), while for the
end B 4 cases: 20, 15, 15, 20.

(1)

where:
n index denoting the current sample of the processed phase
current.
From pre-fault voltages and currents, an active power is
calculated [8]:
Ppre (n ) = 0 . 5[u 1C (n ) i 1C (n ) + u 1 S (n ) i 1 S (n )]

(2)

where:
u1C(n), i1C(n), u1S(n), i1S(n) voltages and currents after fullcycle Furrier filtering (a pair of sine and cosine filters).
The calculated pre-fault power is used for detecting a power
flow direction. Appropriate setting is applied according to the
detected direction.
The calculated samples of the phase shift undergo
fuzzification. For this purpose the set of five consecutive
samples are taken for determining the membership function in
the shape of the triangle. The points of this triangle are
determined with the minimum, average and maximum values
of the phase shift.
Membership function ( ) and setting

Setting

0.8

0.6

F
0.4

F1

0.2

0
132

134

136

138

Phase shift: [deg]

Figure. 3. Illustration of fuzzy comparison.

139

The setting applied in the fuzzy comparison (Fig. 3) was

determined arbitrary, for each fault type separately. In further
investigations, the self-adjusting fuzzy settings will be applied

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
301

Equivalent system
at terminal A
(=0)

Z1SA

(0.656+j7.5)

Z 0SA

(1.167+j11.25)

Equivalent system
at terminal B
(= 15)

Z1SB

(1.31+j15)

Z 0 SB

(2.33+j26.6)

'
Z1L

(0.028+j0.315) /km

Z '0L

(0.275+j1.027) /km

'
C1L

13.0 nF/km

'
C 0L

8.5 nF/km

Series capacitors

0.70 X 1L

Series
compensation
MOV
characteristic:

V
i MOV = P V
VREF

Position of
the compensating bank
P
VREF

0.5 p.u.
1 kA
150 kV

23

Line length
System voltage

160

FA

300 km
400 kV

120

FB
80

40

The presented algorithm for detecting the fault position (as

in Fig. 2) has been reflected in MATLAB [5] environment. The
sample obtained results are shown in Fig. 4Fig. 5.
The example 1, for which the relay input currents are
presented in Fig. 6, is for the fault FA (occurring in front of the
compensating bank) with the following specifications fault
distance: d=0.2 p.u., fault resistance: RF=5 .
The example 2, for which the relay input currents are
presented in Fig. 7, is for the fault FB (occurring behind the
compensating bank) with the following specifications fault
distance: d=0.8 p.u., fault resistance: RF=5 .
The presented sample results illustrate effectiveness of the
proposed detection algorithm. The results of the overall
evaluation are gathered in Table II.

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

Time [s]

Figure 4. Typical shapes of phase shift: a) fault position in front of

compensating bank (FA), b) fault position behind compensating bank (FB).
1

0.8

Decision signal

Line AB

The phase shift is as the one presented by the curve FA in

Fig. 4.
In contrast, for the example 2 (fault occurring behind the
compensating bank) the currents from the faulted phases are
contaminated with the subharmonic oscillations (Fig. 7). The
phase shift is as the one presented by the curve FB in Fig. 4.
In Fig. 5 the decision signal for the example 1 is shown. The
decision that this is a fault in front of the compensating bank
(fault FA) is reached after 16 ms from the fault inception.

Side A - Phase shift [deg]

TABLE I
BASIC PARAMETERS OF THE TEST TRANSMISSION NETWORK

0.6

tdetect=0.016 s

0.4

0.2

TABLE II
EFFICIENCY OF DETECTION OF FAULTS FA, FB
Fault
type
Phasetophase
Double
phaseto- earth
Three
phase

Number
of fault
cases

Detection
of faults FA
[%]

Detection
of faults FB
[%]

Average time
of detection
[ms]

10200

100

100

16

10200

100

100

17

10200

100

100

15

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

Time [s]

Figure 5. Example 1 making a decision with utilizing the relationship (3).

FA fault in front of SCs/MOVs

FB fault behind SCs/MOVs

In Fig. 6 (Example 1) the phase currents under ab fault

occurring in front of the compensating bank are shown. DC
components are present in the currents from the faulted phases.

For each fault type (phase-to-phase, double phase-to-earth

and three phase faults) 30 600 cases were applied in evaluation
of the fault position detecting algorithm.
All faults occurring in front of the compensating bank (fault
FA) and occurring behind the compensating bank (fault FB)
were successfully detected (100% efficiency). This concerns
only inter-phase faults since in the case of single phase faults
the recognition of fault position appeared as unreliable and thus
these results are not shown here.
Average time of the fault position detection is around 16 ms.
In order to get higher speed of detection, the other filtering

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
302

digital algorithm, as for example applying the half-cycle dc

rejection [8] can be applied.

Side A - Phase Currents [kA]

15

5
a b c
0

-5

-15

0.02

0.04

0.06

0.08
Time [s]

0.1

0.12

0.14

0.16

Figure 6. Example 1 currents for fault in front of SCs&MOVs (fault FA).

5

Side A - Phase Currents [kA]

The presented algorithm for detecting the position of the

fault with respect to the compensating device is based on
exploring the change of the phase shift for the protective relay
input currents. The decision with respect to the fault position is
made using fuzzy logic reasoning.
The delivered algorithm has been tested and evaluated with
the fault data obtained from versatile ATP-EMTP simulations
of faults in the test power network with the 400 kV, 300 km
transmission line. Different specifications of faults and prefault power flows (in total, above 30 thousand cases) have been
considered in the evaluation study.
Application of fuzzy logic appeared as the tool which
allowed obtaining efficient detection of the fault place. Interphase faults occurring behind and in front of SCs&MOVs have
been detected perfectly correct (100%).
It is expected that in future research on the issue of detecting
the faulted section of the series compensated line, a multicriteria algorithm will be developed. It seems that only multicriteria approach can assure completely correct recognition of
the fault position, for very wide range of specifications of
faults and the transmission network parameters. This is so also
since the position of single phase-to earth faults is not well
recognised with the presented algorithm.

a b c

V. REFERENCES

[1]

-1

-3

-5

0.02

0.04

0.06

0.08
Time [s]

0.1

0.12

0.14

0.16

Figure 7. Example 2 currents for fault behind SCs&MOVs (fault FB).

IV. CONCLUSION

B. Bachmann, D. Novosel, D. Hart, Y. Hu and M. M. Saha, Application

of artificial neural networks for series compensated line protection, in
Proc. of the International Conference on Intelligent Systems Applications
to Power Systems, Orlando, FL, pp. 6873, Jan. 28Feb. 02, 1996.
[2] CICGRE SC-34 WG-04, Application guide on protection of complex
transmission network configurations, CIGRE materials, August 1990.
[3] P. K. Dash, S. R. Samantaray and G. Panda, Fault classification and
section identification of an advanced series-compensated transmission
line using support vector machine, IEEE Trans. on Power Del., vol. 22,
no. 1, pp. 6773, Jan. 2007.
[4] H. W. Dommel, Electro-Magnetic Transients Program, BPA, Portland,
Oregon, 1986.
[5] MATLAB, High-performance numeric computation and visualization
software. External interface guide, The Mathworks, Inc., January 1993.
[6] W. Rebizant, A. Wiszniewski and L. Schiel, Acceleration of transformer
differential protection with instantaneous criteria, in Proc. Advanced
Power System Automation and Protection, Jeju, Korea, April 2427,
2007.
[7] E. Rosolowski, J. Izykowski and B. Kasztenny, A new half cycle
adaptive phasor estimator immune to the decaying dc component for
digital protective relaying, in Proc. of North America Power Systems
(NAPS) Conference, University of Waterloo, Canada, October 2324,
2000.
[8] E. Rosolowski, J. Izykowski, B. Kasztenny and M. M. Saha, A new
accurate fault algorithm for series compensated lines, IEEE Trans. on
Power Del., vol. 14, no. 3, pp. 789795, Jan. 1999.
[9] A. Wiszniewski, B. Kasztenny, A multicriteria differential transformer
relay based on fuzzy logic, IEEE Trans. on Power Del., vol. 10, no. 4,
pp. 17861793, Oct. 1995.
[10] A. Wiszniewski, J. Szafran, Algorytmy pomiarowe i decyzyjne cyfrowej
automatyki elektroenergetycznej, (Measuring algorithms and decisions
digital power system automation) WNT Warszawa 2001 book in
polish.
[11] Q. Y. Xuan, Y. H. Song, A. T. Johns, R. Morgan and D. Williams,
Performance of an adaptive protection scheme for series compensated
EHV transmission systems using neural networks, Electric Power
System Research, vol. 36, no. 1, pp. 5766, Jan. 1996.

This paper presents the algorithm aimed at recognising the

fault position with respect to the compensating device. The
decision with respect to the fault position, i.e. whether a fault
occurred in front of the bank or behind it is highly required for
design of the adaptive distance relay. Knowledge of the fault
position allows making the distance relay adaptable to presence
of the compensating bank in the fault loop. There is a need for
reflecting the presence of SC&MOV (or SCs&MOVs in the
case of inter-phase faults) in the fault loop, considered by the
distance relay under faults occurring behind the compensating
bank. For this purpose, the differential equation or fundamental
frequency equivalenting approaches, known from the
numerous references, can be applied. By reflecting the
presence of SCs&MOVs in the fault loop (for faults occurring
behind them) the quality of distance protection can be
substantially improved. In particular, one can avoid shortening
of the reach for the high-speed first zone.
Autorzy pragn podzikowa Ministerstwu Nauki i Szkolnictwa Wyszego za finansowanie prac badawczych z zakresu
adaptacyjnych zabezpiecze prowadzonych w ramach projektu badawczego N511 008 32/1688

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
303

Lightning Protection of Aircraft

Lukasz Mackowiak
Wroclaw University of Technology
Wybrzeze Wyspianskiego 27
Wroclaw, Poland
Abstract - this paper concerns lightning phenomenon
and methods of aircraft protection against the lightning effects .

I.

INTRODUCTION

It is estimated that on average, each airplane in the U.S.

commercial fleet is struck lightly by lightning more than once
each year. In fact, aircraft often trigger lightning when flying
through a heavily charged region of a cloud. In these instances,
the lightning flash originates at the airplane and extends away
in opposite directions. Although record keeping is poor,
smaller business and private airplanes are thought to be struck
less frequently because of their small size and because they
often can avoid weather that is conducive to lightning strikes.
The last confirmed commercial plane crash in the U.S. directly
attributed to lightning occurred in 1967, when lightning caused
a catastrophic fuel tank explosion. Since then, much has been
learned about how lightning can affect airplanes. As a result,
protection techniques have improved. Today, airplanes receive
a rigorous set of lightning certification tests to verify the safety
of their designs. Although passengers and crew may see a flash
and hear a loud noise if lightning strikes their plane, nothing
serious should happen because of the careful lightning
protection engineered into the aircraft and its sensitive
components.[6]
II. LIGHTNING PROTECTION REQUIREMENTS
For modern transport aircraft, the ANAC (Civil Aviation
National Agency), FAA (Federal Aviation Administration) and
EASA (European Aviation Safety Agency) include the
Requirements:
25.581 for structural parts lightning protection,
25.901 and 25.903 for engines,
25.954 for fuel system lightning protection,
Also
SAEs
(Society
of
Automotive
Engineers)
Aerospace Recommended Practices 5412, 5413 and 5414 are
extensively used as the basis for the current aircraft lightning
environment definition and for compliance demonstration.
SAE has emitted ARP 5412 regarding lightning environment
and waveforms, and ARP 5413 that provides guidance for
compliance with indirect effects of lightning regulations. In
addition to the 25.1301 and 25.1309 current airworthiness
requirements, the ANAC, FAA and EASA have emitted
25.1316 requirement, which define operation condition for
electrical-electronic system level A (critical system) and level
B /C (essential system).
The requirements of EMI (Electromagnetic Interference) and
lightning direct and indirect effects are complied by

incorporating, since the early design phases, special electrical

bonding and EMI shielding protection devices and techniques:
Metal parts electrical bonding is extensively used;
Composite materials parts include embedded
expanded copper foil and specially developed
electrical bonding interfaces to metal structure;
Wiring harnesses are segregated by separately routing
the different EMI categories of signal and power leads
and shield termination make extensive use of 360
degrees termination and bonding to connector
backshells.
The use of the composite material in fuel tanks and fuel system
are in development, where special attention will be given to
protection against lightning direct effects.[1,3]
III. LIGHTNING PHENOMENON
Lightning is an atmospheric discharge of electricity, which
typically occurs during thunderstorms, and sometimes during
volcanic eruptions or dust storms. In the atmospheric electrical
discharge, a leader of a bolt of lightning can travel at speeds of
60,000 m/s (220,000 km/h), and can reach temperatures
approaching 30000 C. There are some 16 million lightning
storms in the world every year. An average bolt of lightning
carries an electric current of 40 kA, and transfers a charge of
five coulombs and 500 MJ. Large bolts of lightning can carry
up to 120 kA and 350 coulombs. The voltage depends on the
length of the bolt, with the dielectric breakdown of air being
three million volts per metre; this works out to approximately
1GV for a 300 m lightning bolt. With an electric current of 100
kA, this gives a power of 100 terawatts.[7]
The lightning current consists of 4 components (Fig. 1):
Component A appears during initial stroke. The peak
amplitude of this component is equal 200kA 10%
with time duration 500s.
Component B is an intermediate current with average
amplitude about 2kA. The maximum charge transfer
for this component is 10 coulombs.
Component C is a continuing current with amplitude
in the range 200-800A. Charge transfer is equal 200
coulombs 20%.
Component D appears during restrike. The peak
amplitude of this component is equal 100kA 10%
with time duration 500s.
Those current components described above may be utilized for
analyses or test purposes, or for combinations thereof to assess
the potential lightning effects on the aircraft structure or
system, and verify adequacy of protection designs.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
304

Figure 1. Lightning current components [1].

The lightning effects to which aircraft are exposed are divided
into direct effects and indirect effects. Direct effects are
physical effects to the aircraft and/or equipment due to the
direct attachment of the lightning channel and/or conduction of
lightning current and its high power and short time transient.
This includes dielectric puncture, blasting, bending, melting,
burning and vaporization of surfaces and structures. It also
includes directly injected voltages and currents in associated
wiring, plumbing, and other conductive components. The
thermal effects at the arc attachment and the high current
flowing in adjacent structures can produce direct damage. For
example, the heating and mechanical effects of the strike could
result in the puncture of thin fuel tank skins or damage to
control wires. Shock waves may shatter thin panels and the
huge magnetic forces derived from the lightning currents may
buckle metalwork and snap bonding braids. In addition, the
radome, which is a non-conducting aerodynamic fairing over
the weather radar, might be punctured and shattered by a
lightning strike. A further effect resulting from a strike is the
occurrence of sparks which might be generated on the inside of
a fuel tank by the passage of the current. These sparks could
ignite the fuel/air vapour present resulting in fires and
explosions. Indirect effects are resulting predominantly from
the interaction of the electromagnetic fields accompanying
lightning with electrical apparatus in the aircraft. When the
aircraft is struck by lightning, it becomes part of the lightning
channel. Lightning could flow into an attachment point such as
a wing tip and out from another such as a tail fin. A large
current pulse is thus driven through the aircraft enveloping it in
a large and changing magnetic field which can induce voltages
in the long wire and cable runs which are routed throughout the
aircraft. the field penetrates the aircraft by electromagnetic
windows such as the cockpit windshield. Unless protective
measures are taken, induced voltages and currents can easily be
large enough to destroy electrical components or disrupt
electronic and computer systems. It is the same changing fields
which generate the clicks heard on the radio even hundreds of
miles from the lightning source. [1][2]
To analyze of the lightning effect is necessary to determine the
aircraft surfaces, or zones, where lightning strike attachment to
the aircraft is more or less probable.

Figure 2. Example of lightning strike zone details for

Transport Aircraft [3]
The Federal Aviation Administration has defined the following
zones in its Advisory Circular AC 20-53A and in a proposed
draft AC 20-53B associated with specific external lightning
waveforms.
The specific zones definitions are as follows [3]:
Zone 1A: Initial attachment point with low possibility
of lightning channel hang-on.
Zone 1B: Initial attachment point with high possibility
of lightning channel hang-on.
Zone 2A: A swept-stroke zone with low possibility of
lightning channel hang-on.
Zone 2B: A swept-stroke zone with high possibility of
lightning channel hang-on.
Zone 3: Portions of the airframe that lie within or
between the other zones, which may carry substantial
amounts of electrical current by conduction between
areas of direct or swept stroke attachment points.
The advisory circulars do not prescribe the actual locations of
these zones on a particular aircraft since they depend critically
upon the geometry of the aircraft, materials and operational
factors. However, they give generic examples for different
classes of aircraft (large jets, small single and twin-engine
propeller aircraft, helicopters).

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
305

IV. LIGHTNING PROTECTION

It is true that some aircraft are less prone to lightning strikes.
Size, shape, and speed are all aircraft-specific variables which
determine an aircrafts susceptibility to a lightning strike.
However, it is also true that all aircraft damage varies with
aircraft type. Careful aircraft design can minimize lightning
damage. [4]
Some pilots are better at avoiding lightning strikes then
others, but all of them have to remember some basic rules
about flying. Those rules can help to avoid the lightning strikes.
The most important thing is to stay clear of
thunderstorms. Do not attempt to pick your way
through.
The higher the aircraft altitude, the farther away from
a thunderstorm you should fly. Lightning strikes have
been known to occur in the clear air up to 80km
downwind from the nearest thunderstorm.
At low levels, avoid flying close to high surface
features (ridge tops, towers, etc.), or between such
features and an overhead thunderstorm.
If you fly above the freezing level in or near
thunderstorms, you can trigger an in-cloud or cloudto-cloud discharge. If you fly below the freezing level,
you could be involved with a cloud-to-ground
lightning strike. Overall, if you must penetrate or fly
close to a thunderstorm system, you can expect more
strikes penetrating a thunderstorm area well above the
freezing level.
Lightning damage is usually worse for large total
current transfers. At altitudes above the freezing level,
you are more likely to experience longer-lasting
lightning attachments made up of numerous small
pulses and large total current transfer. Below the
freezing level, you are more likely to experience
shorter lightning attachments with a few strong
current pulses; however the total current transfer is
usually less then that above the freezing level.
Electrical activity generated by a thunderstorm may
exist even after the thunderstorm cell has destroyed.
Therefore, avoid penetrating the cirrus decks that
were recently associated with thunderstorms.[4]
The complying of above rules can be treated as a first level of
aircraft lightning protection. If the pilots remember above rules
the probability of lightning strike to the aircraft will be lower.
Unfortunately, this probability will never be equal zero,
therefore we have to apply some protection system or devices
against direct and indirect effects of lightning strikes. The most
popular protection system against lightning strikes is Lightning
Diversion Strips LDS. Lightning Diversion System has
developed a new and improved lightning protection device that
diverts lightning strokes from aircraft nose radomes and other
sensitive areas. The protection system consists of a segmented
diverter strip which provides maximum multiple-strike
protection.

Figure 3. Application of LDS [8]

Attached to an aircraft's radome, the system allows a lightning
stoke to travel safely and directly to ground in an ionized
channel created in the air above the diverter strip. It combines
permanent protection with a low drag aerodynamics and has
insignificant effect on radar antenna radiation patterns. The
electrostatic shield created by the system provides a new
source of streamers outside the radome wall to the fuselage.
The small diameter of the disc segments makes the strip
compatible with radar system operation. If necessary, disc size
can be reduced for optimum antenna patterns at higher
frequencies.[8]
To protect an electrical devices we can apply surge suppressors.
For example the 115 VAC high energy surge suppressor
combines in one unit the low clamping voltage and high surge
current capabilities needed to protect solid state power and
electronic systems from severe lightning-induced and other
over-voltage surges. This suppressor meets the need for low
clamping voltage and exceptionally high surge current
capability packaged in a single, small size unit. The suppressor
is packaged to withstand rugged airborne and ground-based
environments and will clamp repetitive high energy transients.
The 115 VAC high energy surge suppressor is especially
suitable for clamping lightning-induced surges appearing on
aircraft 115 volt AC power distribution busses feeding solidstate avionics. The suppressor clamps the high energy surges to
levels easily tolerated by most solid state power supply inputs
without interruption of power or tripping of circuit breakers. Its
surge current capability exceeds that of many spark-gap
devices, and its response (turn on) time is virtually

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
306

instantaneous. The suppressor meets applicable requirements

of most aircraft power quality specifications. [8]
Applications of surge suppressors in aircrafts:
Primary power distribution busses.
Secondary power distribution circuits, at utilization
equipment.
Externally mounted electrical apparatus (windshield
and air data probe heater circuits, navigation lights,
and electrical de-icing systems.)
AC power and control circuits within advanced
composite airframes
V. CONCLUSIONS
According to [5], lightning aircraft incidents most often
occur when the air temperature is in the range
of -5 to 0C. Lightning incidents frequency depend on altitude
and season, thus in winter more incidents happen at about 1km
above sea level and in summer more incidents occur at about
5km above sea level. Taking into account that pilots know this
information and rules from chapter IV, flying by plane should
be very safe. And it is. But sometimes thunderstorm can
surprise even the best of pilot. Therefore all kinds of airships
have to fulfill a lot of lightning protection requirements. But
we have to remember, that even the best lightning protection
system does not replace the human being. Thus, in my opinion
if the pilot sees a thunderstorm he should change flight
direction to bypass this hazard.
REFERENCES
[1]

[2]
[3]

[4]
[5]
[6]
[7]
[8]

S. A. Baldacim, N. Cristofani, J.L.F. Junior, J.R. Lautenschlager,

Lightning effects in aircraft of the composite material, 17 CBECIMat Congresso Brasileiro de Engenharia e Cincia dos Materiais, 15 a 19 de
Novembro de 2006, Foz do Iguau, PR, Brasil
John Hardwick Protect and survive, Aircraft Engineering and
Aerospace Technology, Volume 69 Number 5 1997 428431
Alain Broc, Philippe Lalande, Emmanuel Montreuil, Jean-Patrick
Moreau, A lightning swept stroke model: A valuable tool to investigate
the lightning strike to aircraft, Aerospace Science and Technology 10
(2006) 700708
Anonymous, New research produces new information about aircraft and
lightning strikes, Flying Safety; Aug 1997; 53, 8; Career and Technical
Education, pg. 29.
Rakov, V. A., Uman, M. A., Lightning: Physics and Effects,
Cambridge University Press, UK 2003.
[Link]
[Link]
[Link]

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
307

Elimination of Chaotic Ferroresonance in power transformers including

Nonlinear Core Losses applying of Neutral Resistance
[Link]

[Link]

abbasi@[Link] Rostami@[Link]

[Link]
[Link]@[Link]

[Link]
[Link]@[Link]

Faculty of Engineering,
Shahed University,
Khalig-e-Fars free way,
1417953836 Tehran, Iran,
P.O. Box: 18151/159,
Tel: (+98-21)66639821, Fax: (+98-21)51212105, Tehran, Iran
Abstract
This paper studies the effect of neutral
resistance on the onset of chaotic
ferroresonance and chaotic transient in a
power transformer including nonlinear core
losses. It is expected that this resistance cause
ferroresonance dropout'. Time-domain study
has been carried out to study this effect.
Simulation has been done on a three phase
power transformer rated 50 MVA, 635.1 kV
with one open phase. The magnetization
characteristic of the transformer is modeled by
a single-value two-terms polynomial. The core
loss is modeled by third order power series in
terms of voltage includes nonlinearities in core
[Link] simulation results reveal that
connecting the neutral resistance to the
transformer, exhibits a great mitigating effect
on voltage ferroresonance. Phase plane along
with bifurcation diagrams are also derived.
Having Significant effect on the onset of
chaos, the range of parameter values that may
lead to chaos along with ferroresonance
voltages has been obtained and presented.
Key word: Chaos, Bifurcation diagram
,Ferroresonance ,Power transformer ,Neutral
Resistance
1. Introduction
Ferroresonance may be initiated by switching
actions, or load shedding in power system. It
can produce unpredictable over voltages and
high
currents.
The
prerequisite
for
ferroresonance is a circuit containing nonlinear
iron core inductance and a capacitance. Such a
circuit is characterized by simultaneous
existence of several steady-state solutions for a
given set of circuit parameters. The abrupt

transition or jump from one state to another is

triggered by a disturbance, switching action or
a gradual change in system parameters.
Typical cases of ferroresonance are reported
in. [1], [4]. Theory of nonlinear dynamics has
been found to provide deeper insight into the
phenomenon [5].
Fundamental theory of nonlinear dynamics
and chaos can be found in [4], [7]. The
susceptibility of a ferroresonance circuit to a
quasiperiodic
and
frequency
locked
oscillations are presented in [8], [9], The effect
of initial conditions which is investigated. [10]
is a milestone contribution
highlighted the effect of transformer modeling
on the predicted ferroresonance oscillations.
Using a linear model, [11] has shown the
effect of core loss in damping ferroresonance
oscillations. Reference [12] emphasizes
treating core loss as a nonlinear function of
voltage. An algorithm for calculating core
losses based on no-load characteristics of
transformer is given in [13].The mitigating
effect of transformer connected in parallel to a
MOV arrester is illustrated in [14].Effect of
circuit breaker grading capacitance on
ferroresonance in VT is investigated in [15]. In
all previous studies, the effects of neutral
resistance including nonlinear core losses are
neglected. In these works, it has been shown
that nonlinear core loss can decrease
ferroresonance
phenomena
in
power
transformer . Later, the effect of neutral
resistance on power transformer including
linear core loss has been investigated [16].
Current paper studies the effect of neutral

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
308

resistance on the global behavior of a

ferroresonance circuit with nonlinear core loss.
1. System Modeling (Base System)
System Modelling Without Neutral Resistance
in Transformer is assumed to be connected to
the power system while one of the three circuit
breaker poles is open and only two phases of it
are energized, which produces induced voltage
in the open phase. This voltage, back feeds the
line, and consequently ferroresonance will
occur if the distribution line is enough
capacitive. System involves the nonlinear
magnetizing reactance of the transformer's
open phase and resulted shunt and series
capacitance of the distribution line.

Fig. 2. Nonlinear characteristics of transformer

core with different values of q
Beside,to study the case, the core loss model is
formulated by a third order power series whith
coefficients, fitted to match the hysteresis and
eddy current nonlinear characteristics given in
[1]:
2

i Rm = h 0 + h1 v m + h 2 v m + h 3 v m
Per unit value of

(2)

is given in (3)

i Rm = .000001 + .0047Vm .0073Vm + .0039Vm (3)

2

Fig 1. Equivalent circuit of system

Fig.1. shows the equivalent circuit of system
described above. The magnetization branch is
modelled by a nonlinear inductance in parallel
with a nonlinear resistance accordingly
represent
the
nonlinear
saturation
characteristic
and nonlinear hysteresis
,
and eddy current characteristics
respectively. The hysteresis and eddy current
characteristics of core can be obtained from
no-load statues of transformer by applying the
algorithm given in [13]. The iron core
saturation cure characteristic is given by:
(1)
Exponent q depends on the degree of
saturation. It has been found that for adequate
representation of the saturation characteristics
of a power transformer the exponent q may
acquire values 5, 7, and 11. Fig.2 shows iron
core characteristic for q=5, 7, 11.

The differential equation for the circuit in fig.

1 neglecting neutral resistance can be
presented as follows:

d
( 4)
dt
1
V&c = (h0 + h1V + h2V 2 + h3V 3 + a + bq )(5)
C
dE V&c V&L = 0
(6)
V=

dE

1
(h0 + h1V + h2V 2 + h3V 3 + a + bq )
C

d 2
= 2 (7 )
dt
Where E is the peak value of the voltage
source, shown in fig.1.
Typical values for system parameters without
neutral resistance are as follows:

for q = 11 s1 = .0067 , s 2 = .0001

for q = 7 s1 = .0067 , s 2 = .001

for q = 5 s1 = .0071 s 2 = .0034

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
309

 = 377 rad/sec

Phase Plane Diagram for q=7 Without Neutral Resistance

R = 8067
C pu = 0.7955
Vbase = 635.1kv , I base = 78.72 A
And initial conditions are:
0

0.0;

p 0

1.1. Simulation Results

Table 1 shows different values of E,
considered for analyzing the circuit in absence
of neutral resistance.

Fig 4. Phase Plane diagram of system with

q=7, without neutral resistance

Phase Plane
Diagram
for q=11Without
Neutral Resistance
Phase
Plan Diagram
for q=11 Without
Neutral Resistance

Table 1: Simulation results (Without Neutral

Resistance)
q/ 1 3
5
7
9
10
E
5

P P2 P3
2
7
P P2 P3
2
11 P P3 chaoti
2
c
*Pi: Period i

P6
chaotic
chaotic

Chaot
ic
Chaot
ic
P4

P9
Chaot
ic
P4

Fig 5. Phase Plane diagram of system with

q=11, without neutral resistance

Fig.3, 4, and 5 show the phase plane plot of

system states without neutral resistance For
E=3 p.u. Figs.6, 7 and 8 show the
corresponding bifurcation diagrams of system
states without neutral resistance and including
nonlinear core losses for q=5, 7, 11 which
depicts chaotic behaviour.
Phase Plane Diagram for q=5 Without Neutral Resistance

Fig 6. Bifurcation diagram with q=5, without

neutral resistance

Fig. 3. Phase Plane diagram of system with

q=5, without neutral resistance
8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
310

been added to the system and its value is given

below:

R neutral = 40k
The differential equation for the circuit in fig.9
can be presented as follows:
dE V&c V&Rneutral V&L = 0

1
(h0 + h1V + h2V 2 + h3V 3 + a + bq )
C
d 2 d
d 2 d2
d 2
RNeutral(+h1 2 + 2h2 2
+ 3h3 2 (
)+
dt
dt dt
dt
dt
d
d
d 2
a + qb q1 ) = 2
dt
dt
dt
Where is the flux linkage and V, is the
voltage of transformer
Initial conditions:
0
0
0
2
dE

Fig 7. Bifurcation diagram with q=7, without

neutral resistance

2.1. Simulation results

Fig 8. Bifurcation diagram with q=11, without

neutral resistance

2. System modeling with neutral resistance

Including Neutral resistance,system of fig 1 ,
can be Modyfied as shown in fig 9.

Fig.10 shows the corresponding phase plane

diagram and Fig .11 shows voltage wave form
for system in fig.9 in case of q=11.
Consequently figs.12, 13 and 14 shows
bifurcation diagrams for corresponding system
including neutral resistance. It can be seen that
chaotic region mitigates by applying neutral
resistance and there is no chaotic region. wich
is means, tendency to chaoticbehaviour
reducesdramaticallyif neutral resistance effect
been included.
Comparing to table 1 table 2 includes the
results
considering circuit
with neutral
resistance.
Again
,
its
clear
that
ferroresonance drop will be occurred.

Table 2: Simulation results (with neutral

resistance)
q/E
1
2
3
4
5
6
Fig. 9. Equivalent circuit of system with
neutral resistance
Typical values for various system parameters
has been considered for simulation kept the
same by case 1, while neutral resistance has

P1

P1

P1

P1

P2

P2

P1

P1

P2

P3

P3

P3

11

P1

P1

P3

P3

P3

P3

*Pi: Period i

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
311

Phase Plan Diagram for q=11 With Neutral Resistance

Fig 10. Phase Plan diagram of system with

q=11, including neutral resistance

Fig 11. Time domain simulation with q=11,

including neutral resistance

Fig 13. Bifurcation diagram with q=7, with

neutral resistance

Fig 14. Bifurcation diagram with q=11, with

neutral resistance

3. Conclusions

Fig 12. Bifurcation diagram with q=5, with

neutral resistance

The dynamic behaviour of a transformer can

be characterized by different presentation.
Inclusion of nonlinearity in the core loss
reveals that bifurcation diagrams will be
smoother but more complicated when
compared with linear models.
Adding Neutral resistance , It has been shown
that system has been greatly affected by it. The
presence of the neutral resistance results
clamping the Ferroresonance over voltages in
studied system. The neutral resistance
successfully, suppresses the chaotic behavior
of proposed model. Consequently, the system
shows less sensitivity to initial conditions in
the presence of the neutral resistance.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
312

4. References
[1] H.W. Dommel, A. Yan, R.J.O. De
Marcano, A.B. Miliani, in:H.P. Khincha(Ed.),
Tutorial Course on Digital Simulation of
Transients in Power Systems (Chapter 14),
IISc, Bangalore, 1983,pp. 17_/38.
[2] E.J. Dolan, D.A. Gillies, E.W. Kimbark,
Ferroresonance in a transformer switched with
an EVH line, IEEE Transactions on Power
Apparatus and Systems PAS-91 (1972)
1273_/1280.
[3] R.P. Aggarwal, M.S. Saxena, B.S. Sharma,
S. Kumer, S. Krishan, Failure of
electromagnetic voltage transformer due to
sustained overvoltage on switching*/an indepth field investigation and analytical study,
IEEE Transactions on Power Apparatus and
Systems PAS-100 (1981) 4448_/4455.
[4] C. Kieny, Application of the bifurcation
theory in studying and understanding the
global behavior of a ferroresonant electric
power circuit, IEEE Transactions on Power
Delivery 6 (1991)866_/872.
[5] A.E. Araujo, A.C. Soudack, J.R. Marti,
Ferroresonance in power systems: chaotic
behaviour, IEE Proceedings-C 140 (1993)
237_/240.
[6] S. Mozaffari, S. Henschel, A.C. Soudack,
Chaotic ferroresonance in power transformers,
IEE Proceedings*/Generation Transmission
and Distribution 142 (1995) 247_/250.
[7] B.A. Mork, D.L. Stuehm, Application of
nonlinear
dynamics
and
chaos
to
ferroresonance in distribution systems, IEEE
Transactions on Power Delivery 9 (1994)
1009_/1017.
[8] S.K. Chkravarthy, C.V. Nayar, Frequencylocked and quasi periodic (QP) oscillations in
power systems, IEEE Transactions on Power
Delivery 13 (1997) 560_/569.
[9] S. Mozaffari, M. Sameti, A.C. Soudack,
Effect of initial conditions on chaotic
ferroresonance in power transformers, IEE
Proceedings*/Generation, Transmission and
Distribution 144(1997) 456_/460.
[10] B.A. Mork, Five-legged wound*/core
transformer model: derivation, parameters,
implementation, and evaluation, IEEE
Transactions on Power Delivery 14 (1999)
1519_/1526.
[11] B.A.T. Al Zahawi, Z. Emin, Y.K. Tong,
Chaos in ferroresonant wound voltage
transformers: effect of core losses and
universal
circuit
behavioral,
IEE

Proceedings*/Sci. [Link]. 145(1998)

39_/43.
[12] IEEE Working Group on Modeling and
Analysis of Systems Transients, M.R. Iravani,
Chair, Modeling and analysis guidelines for
slow transients*/part III: the study of
ferroresonance, IEEE Transactions on Power
Delivery, 15 (2000) 255_/265.
[13] W.L.A. Neves, H. Dommel, on modeling
iron core nonlinearities, IEEE Transactions on
Power Systems 8 (1993) 417_/425.
[14] K. Al-Anbarri, R. Ramanujam, T.
Keerthiga, K. Kuppusamy,Analysis of
nonlinear phenomena in MOV connected
Transformers,IEE
Proceedings*/Generation
Transmission and Distribution148 (2001)
562_/566.
[15]: [Link], [Link],[Link];
Effect of Circuit Breaker Shunt Resistance on
Ferroresonance phenomena in Voltage
Transformer,
IEEE
2009
conference,
Georgia,USA
[16] [Link],[Link] [Link].
Analysis of Ferroresonance Phenomena in
Power Transformers Including Neutral
Resistance Effect, IEEE 2009 conference,
Georgia,USA

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
313

Geothermal energy as renewable energy source

S. Kitscha, J. Schmeckebier, M. Steckel
BTU Cottbus
Cottbus, Germany

I. INTRODUCTION
The earth is a dynamic planet and can be described as a
gigantic thermal engine. The available heat is called
geothermal energy.1 Geothermal energy is a renewable energy
source just like wind, hydro or solar power. By using
geothermal energy, greenhouse gas emissions can be reduced
significantly. However, other than solar or wind power,
geothermal energy is neither dependent on climatic influences,
nor daytime or season and is available at almost any place and
time. Although as a renewable energy source geothermal
energy offers many advantages and is available at many
potential sites, it has so only a comparatively low proportion of
the total energy supply so far. As experts assume, geothermal
energy consists of approximately 30% residual heat caused by
the earths formation process about 5 billion years ago, when
gases, rocks and dust were condensed. In this process a huge
amount of energy was released. Temperatures of about 6000
C are reached in the earths core.2
II. APPLICATION
Geothermal energy is generally subdivided into depth and
near-surface geothermal energy. The near-surface geothermal
energy is mainly used to heat or cooling buildings. However,
the depth-geothermal energy primary generates electricity and
heat in (for) power plants, since this method is too complex to
be used for cooling or heating individual buildings. Depthgeothermal energy, based on its technical application, is further
subdivided into hydrothermal geothermal energy, HDR
systems and deep geothermal heat probes. Geothermal energy
enables a supply of individual homes, mostly detached houses,
independently from the main power supply network.3 In order
to realize this, there are many different techniques available
today, from which private households can to select the most
effective according to their specific situation and region.
The heat pump is the heart of each individual plant. This heat
pump receives the heat transfer medium that was headed by the
ground. Then, the heat pump absorbs the heat and feeds the

storage tank from which the houses entire heating and hot
water system is supplied. For security reasons, an additional
boiler can be installed in the system. This is necessary because
it might happen that the geothermal heat collectors are
influenced by the seasons, so that during winters the energy
from the ground is not sufficient in order to heat the building.
The geothermal heat collectors are placed horizontally at 80 160 cm under the surface to be subject to weather conditions.4
The most widespread plant is the geothermal heat probe. When
installing the geothermal heat probe, first a plastic tube bundle
is inserted in a prepared hole. Then the ring area is sealed with
a bentonite / cement / water mixture. The Plastic tube bundle
contains a heat carrying liquid, usually water with an
antifreeze. They are inserted 25 to 200 meters vertically into
the floor and deliver a constant temperature level. Because a
legal procedure is necessary for drilling holes into the soil
deeper than 100m, most of these facilities do not exceed this
limit. The heat transfer medium circulates inside the heat
probes, continuously receiving heat from the ground and
transporting it up to the heat pump. This method does not only
provide heat (and energy) for individual houses but also
supplies offices and commercial buildings, even apartment
complexes or whole neighborhoods.5 Energy earthbound piles
and static parts are needed for new buildings anyway. These
parts can be equipped with heat exchanger tubes, so that
combined with a heat pump buildings can be economically
heated and cooled.
III. INVESTMENTS
Regarding the investments for a geothermal plant there are five
points to consider. In the calculation the costs for official
approvals, drilling (only for geothermal heat probes), the heat
pump, a separate electric meter and several accessories need to
be taken into account. The investments for these facilities are
about 15600-25150. Thus the costs for geothermal power
plants are about 3000-5000 higher than for gas- or oilheaters. Since the investments for geothermal power plants are
comparatively (quite) high, the government offers financial
help to

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
314

Activity
Official approval
Drilling
Heat pump (to 17 kW)
Accessories
Electric meter
Table 1:

Costs
500 1,500
60 - 90 per meter
9,000 13,000
500 1,500
500 1,500

Overview about acquisition costs (2007)

([Link]
kosten/Standardpage_7.html)

support the installations of environmentally friendly

equipment.6 Since 1.1.2008 the installation of heat pumps in
the residential and commercial sector has been supported by
the federal government.
IV. POTENTIAL
The geothermal energy offers a big potential as an alternative
energy source for the future. To reach an exact appraisal of this
form of power generation it is necessary to look at the energypolitical aims. These aims concern the economic efficiency,
environmental compatibility and last but not least the security
of the energy supply. Concerning the environmental
compatibility, the emission of greenhouse gases of alternative
energy sources should be as low as possible which helps to
protect the climate and environment with a friendly energy
supply. The security of energy supply should be improved by
the high-priority use of locally available ressources. The
steadily rising energy prices are in the focus of the economic
efficiency and the alternative energy source should allow an
increasingly economic use and offer independence from high
energy expenses.7
In the following it is demonstrated to what extend the
geothermal energy, as an alternative energy source, fulfill the
energy-political objectives. A cost saving potential of
approximately 80% was identified in comparison to fossil
energy sources by the use of the geothermal energy.
Furthermore the geothermal heating installation needs only
small quantities in electric energy in comparison to gas heating
installations or oil heating installations. It has to be mentioned
here, that the effectiveness of this technology is a multiple
higher than with fossil heating systems. A further advantage of
these systems is that the whole installation could not only be
used for heating in winter, but also for cooling in summer.
Because of this advantage the user does not have to pay for an
additional cooling system. Moreover the costs of maintenance
are less expensive than for traditional heating systems. After
some time it is only necessary to substitute filters of the heat
pump and gaskets of the whole cycle system. Geothermal
installations can be also installed in any building to any time

and they have a life span of approximately 50 years. However,

negative points are the high costs and requirements for the
building insulation. The requirements for the buildings
insulation are so high because geothermal systems heat a
building substantially slower than customary systems and take
a lot of time when the building has an insufficient insulation.
Finally it is to ascertained that this form of power production is
absolutely economic.8
When you take a look to the environmental compatibility, a
representation of the environmental consequences of the
geothermal energy must follow. First, there should not be
released any CO2-emissions by the use of geothermal energy.
The only component which needs electric energy is the heat
pump. Because of this, the usage of this energy form is directly
connected to the CO2-intensive generation of electricity.
Moreover share-fluoridated hydro-fluorocarbons (HFC) are
used in heat pumps. These media do not attack the ozonosphere;
nevertheless, they forward the greenhouse effect. If these HFCs
reach the atmosphere the influence to the greenhouse effect is
approximately 1500-3000 stronger than CO2-emissions.9
Another disadvantageous is that with very low outside
temperatures an additional heating system like gas- or oil
heating is needed because the energy from the earth is not
sufficient. There are often problems with the geothermal
energy collectors, because these lie close under the earth's
surface and can only record a relatively little warming up. At
last it is ascertained that the geothermal systems are absolutely
environment-friendly as an alternative energy source, but there
is still a big need for action concerning the used cooling
medium.
The security of the energy supply is for all systems guaranteed,
when the geothermal energy is only used privately. Then there
are no regional restrictions. Indeed, it is to be noticed, that at
some places the geothermal energy cannot be used, even if the
geologic conditions would be given. This is justified upon to
the fact that geothermal energy probes and geothermal energy
collectors needs a lot of space. A surface which is one and a
half times larger as the surface to be heated is an approximate
value. This is especially problematic for households which are
based on a relatively small property. These households could
install only smaller variations of this technology in the earth
and use them merely for heating or cooling support which is
economically questionable. So they have to abdicate of this
alternative energy source. The application ranges of depthgeothermal energy are more limited than near-surface
geothermal energy. This energy system type can be used for
electricity and heat generation in power plants. But the

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
315

geothermal requirements must be given. This is only possible

at places where the points of the tectonic disks are. Otherwise
the bores are too deep in the earth and a cost-covering
production is impossible. In summary geothermal energy is
theoretically available everywhere, but in many areas there are
restrictions for the use in private households.10

illustration 2 shows. World market leaders in the area of direct

use will be China, Sweden, the USA, Turkey, Iceland and
Japan. Leading and at the same time the largest growth market
is the People's Republic of China at this energy form. But
Iceland is worldwide at first place with 25% of the electricity
and 85% of the required space heating made out of geothermal
energy.11

IV. SUMMARY
About 0.4% of the worldwide power generation is geothermal
energy. The annual growth rate was stable in the last decade
and is still at about 5.0% (about 200-250 MW capacity increase
p.a.). Illustration 1 shows the previous development and the
future trend in comparison to other energy forms. The
geothermal capacity installed worldwide will exceed 10 GW
(year 2008) soon, for 2010 the forecast is about 11 GW. Within
the next 30 years an increase to about 85 GW is regarded as
possible. The most important markets for this energy form are
the USA, Philippines, Mexico, Indonesia, Japan, New Zealand,
Italy and Iceland. Iceland is advancing the development of the
geothermal capacity and the research very actively at present.

Illustration 2: Quota of the relevant energy sectors of the worldwide provision

to geothermal using warmth (in MTOE) for 2004 and forecast
for 2030 to IEA 2007
([Link]
/papers/3D_4_PONWEISER_K_P.pdf)

Illustration 1: Electricity generation (in TWh/a)-temporal development of the

different energy sectors to 2005 and forecast to 2030
([Link]
/papers/3D_4_PONWEISER_K_P.pdf)

The biomass still dominates globally the field of the warmth

generation. By the increase in the world population the need
for heated rooms and energy supply for the eating preparation
also rises further strongly. For this 58% of renewable energies
are used. The renewable energies which fundamentally
contribute to the supply of useful heat are solar energy and
geothermal energy. The capacity of the geothermal energy
installed at present is only a third in comparison to the solar
heat because of the higher degree of utilization the energy
output is comparable. An increase for these energy forms
around 10-fold at solar energy utilization and 6-fold at
geothermal energy is forecast for the year 2030, how

The geothermal energy offers an enormous rate of potential for

the generation of electricity and heat. The two and a half-fold
amount of energy which the mankind needs rises from the
inside of the earth daily. The geothermal energy is available
every day and season and is independent of weather influences
or climate conditions. The huge potential of this alternative
energy source was recognized meanwhile and utilization was
started. Nowadays geothermal energy gets more and more
frequently in the focus of political discussions about the future
energy supply. This has been the case since the costs for fossil
fuels increase rapidly. Due to the long-term safeguarding of the
geothermal energy as an alternative energy source in
combination with the flexible use possibilities in the areas of
heating, cooling and electricity generation the number of plants
which are planned and installed worldwide, increase steady.
REFERENCES
[1]

Innovationsbeirat der Landesregierung (2003):

Zukunft der Energieversorgung, Springer Verlag,
Heidelberg

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
316

[2]

[3]

[4]

[5]
[6]

Geitmann, S. (2005): Erneuerbare Energien und

alternative Kraftstoffe - Mit neuer Energie in die
Zukunft, 2. Auf., Hydrogeit Verlag, Kremmen
Greenpeace
(2000):
2000MW
sauber!
[Link]
themen/energie /Studie_2000MWsauber.pdf
Umweltministerium Baden Wrttemberg (2008):
Leitfaden zur Nutzung von Erdwrme mit
Erdwrmekollektoren,
1.
Aufl.,
[Link]
[Link]?command=downloadContent&file
name=Leitfaden_Erdwaermekollektoren.pdf
The same as in [5]
[Link]
(2008):
Jetzt
in
Zukunftswrme investieren und der Staat zahlt mir,
[Link]
nt/STIEBELELTRON/themenwelt/informieren/inhalt/
[Link]

[7]

[8]

[9]

[10]

[11]

Institut fr Energetik und Umwelt gGmbH:

Geothermische Strom- und Wrmeerzeugung in
Deutschland und Europa Stand und Perspektiven,
[Link]
[Link]
(2008):
Geothermische Heizung Vorteile und Nachteile,
[Link]
sche_ heizung_Vorteile_und_ [Link]
WDR (2008): Nicht alle Wrmepumpen retten das
Klima
[Link]
che/energie/regenerative_energien/interview_waerme
[Link]
[Link] : Vor- und Nachteile der Geothermie,
[Link]
6. Internationale Energiewirtschaftstagung an der TU
Wien - G02 IEWT 2009: Der Beitrag der Geothermie
[Link]
ers/3D_4_PONWEISER_K_P.pdf

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
317

Slot harmonics in stator current as syptoms

of reparation of induction motors
Jan Pytlarz, Pawe Ewert
Wrocaw University of Technology
Janiszewskiego 8
Wrocaw, Poland
Abstract-Nowadays in many industry branches all over the
world more than 90% of all installed motors are of induction
motor type. During induction motor operation in electrical parts
of stator and rotor circuits (windings) and in mechanical part of
the motor and cooperating machine together with coupling
elements there occur numerous failures. Every failure brings
disturbances in technological process and in worst case leads to
whole plant stoppage, what in consequence leads to rise of
production costs. In this connection early failure detection on the
electrical drives monitoring basis has a great weight. In the paper
the attention will be focused on one of the mechanical failures
eccentricity, that is non-axial rotor to stator orientation. There
will be presented theoretical considerations and examination
results connected with this problem. On the basis of carried out
theoretical and experimental considerations there will be made an
assessment of stator current signal usability in induction motor
operation diagnostics.

I.

INTRODUCTION

The history of failure diagnostics is old as the motor itself.

Initially the electrical motors producers and users, to assure
safe and reliable operation, have turned to simple protection
connected with overcurrents, overvoltages, earth faults, etc.
Nowadays in many industry branches all over the world
more than 90% of all installed motors are of induction motor
type. During induction motor operation in electrical parts of
stator and rotor circuits (windings) and in mechanical part of
the motor and cooperating machine together with coupling
elements there occur numerous failures. Every failure brings
disturbances in technological process, and in worst case leads
to whole plant stoppage, what in consequence leads to rise of
production costs. In this connection early failure detection on
the electrical drives monitoring basis has a great weight.
Therefore to avoid production stoppages, lower the repair
costs and lower the operation costs there should be assured
ways of their detection and information about their occurrence.
Ideal motor monitoring system should recognize all
possible failures kinds which can occur in tested drive. It
should be noticed that simultaneous occurrence of several
failures gives similar, and even the same, quality effects.
Occurrence of such case complicates the estimation of their
quantity influence on the machine condition. Without
independent standard creation for every failure we cannot

estimate the degree in which the particular disturbance affects

the total condition of examined circuit.
II.

INDUCTION MOTOR FAILURE REVIEW

To assure safe and reliable operation of electrical motors

initially the electrical motors producers and users have turned
to simple protection connected with overcurrents, overvoltages,
earth faults, etc. However electrical motors development,
functions performed by them in technological process have
feed to diagnostics development. At present non-planned motor
standstills can lead to large financial losses.
Main failures of electrical motors can be classified [2]:
1) Stator failures, which lead to break or short-circuit of
one or several stator phases,
2) Improper stator winding connection,
3) Rotor rods break or short-circuit ring cracking,
4) Statical or/and dynamical irregularities of air gap,
5) Shaft bend (connected with dynamical eccentricity),
which can lead to friction between stator and rotor
causing serious damage of stator core and windings,
6) Shorted rotor exciting windings,
7) Transmission and bearing failures.
This failures cause among other things changeable size of
the air gap, current asymmetry, pulsating increase of torque,
lowering of mean torque, loss increase, efficiency lowering and
heat excess.
To detect failures there have been elaborated many
diagnostics method. These methods take advantage of
achievements form different science [2]:
1) Methods based on stator current spectrum analysis
(MCSA motor current signature analysis),
2) Methods based on measurements of engine noise and
vibration,
3) Methods based on flux analysis (additional coils on
stator shaft winding),
4) Methods based on artificial intelligence models and
neural networks,
5) Methods based on temperature measurement,
6) Methods based on chemical analysis,
7) Radio frequency emissions monitoring,

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
318

8) Methods based on infrared radiation recognition.

III.

ECCENTRICITY AS FAILURE SOURCE

According to Nandis [2] motor eccentricity is a state of

unequal air-gap which exists between the stator and rotor.
Eccentricity detection and diagnosis is a very important
element of induction motor condition monitoring. It is caused
by very small air-gap toleration between stator and rotor. Small
tolerance limit exceed can lead to serious machine failure
(stator friction against rotor which in consequence leads to
stator or rotor failure), but also deepen the emergency state,
caused by other adverse phenomena, such as: asymmetrical
feeding, stator or rotor circuit damage, work with excessive
load, etc.
Air-gap eccentricity can be a primary or secondary failure.
We are speaking about primary failure when the motor has the
asymmetry from the operation beginning. It is caused most
often by oval stator hole or when the rotor cross-section is not
a circle. While when as consequence of sustained non-axial
operation the asymmetry of air-gap will exist, then we speak
about secondary failure.
We distinguish three types of eccentricity:
static eccentricity,
dynamical eccentricity,
mixed eccentricity (simultaneous occurrence of static
and dynamical eccentricity).
In new motors the allowable air-gap eccentricity is up to
10% [2]. Although, the producers try to assure as small as
possible total eccentricity level to reduce the vibrations and
noise, and to lower the asymmetric radial force.
Most often in motors the mixed eccentricity occurs.
In this case the rotation axle can be found between C1 and C2
as shown on figure 1.

detected and controlled because of its deepen tendency, what

in consequence can lead to motor permanent failure.
Eccentricity monitoring and diagnostics causes a lot of
trouble because it must be realized during motor normal
operating conditions, non-invasive not to change the motor
forces. For this reason, except monitoring using X-rays which
is very costly, most often applied method is the spectral
analysis of stator current.
IV.

SLOT

HARMONICS IN STATOR CURRENT SPECTRUM

Induction motor rotor out-ofline location in respect to

stator brings air-gap asymmetry. Because of characteristic
changes of this asymmetry in mutual magnetic couplings
between motor windings influences on stator current spectrum
form. That is the reason for which the method of stator current
spectral analysis (MCSA) is the most often used in eccentricity
detection [1].
In case of induction motors characteristic frequencies
describing every kind of eccentricity can be written by
equation:

1 s
f e = f s ( kN r n d )
nw
pb

(4.1)

where:
f s net frequency, k = 1,2,3,... , every integer,
quantity of rotor slots,

Nr

n d = 0 in case of static eccentricity,

nd = 1,2,3,...
in case of dynamical eccentricity ( n d dynamical eccentricity
grade),

slip,

pb

number

of

pole

pairs,

n w = 1, 3, 5, 7,... grade of stator temporal harmonics.

On the ground of dependence (4.1) in stator current there
can be determined the frequencies connected with static and
dynamical eccentricity and so called principal slot harmonics
(PSH). In a healthy motor those harmonics can be used to
estimate the angular speed.
Characteristic frequencies connected with static eccentricity
are described by the equation:

f es = f s k r (1 s ) 1
pb

Figure 1. Coaxial and eccentric location of rotor in stator hole

Eccentricity occurrence often does not exclude the motor
from further operation. In this case the eccentricity should be

(4.2)

where k = 0, 1, 2,....
In case of simultaneous occurrence of static and dynamical
eccentricity harmonics with frequencies positioned near net
frequency occurs:

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
319

f ed = f s kf r

(4.3)
(4.3)

Frequencies described with following dependencies occur

in all motors. Low frequencies components cause rising of high
frequency components determined by equation (4.1). [1]

5. LABORATORY TESTS
The tests were made using Brael&Kjaer PULSE 3560
analyzer. For INDUKTA SSh 90L-4 engine with parameters:
Nr=26,
pb=2,
and work with nominal load (s=0,0366) we get, basing on
equation (4.1), the following harmonics distribution:

6. FINAL THOUGHTS
Basing on above mentioned considerations and mentioned
examples we can formulate the following conclusions:
Stator current spectral analysis a great method to
monitor, diagnose and valuate of basic failures
occurring in inductive motor;
Detailed analysis is possible with assurance of high
resolution of measuring apparatus and wide frequency
band in measured signal;
Spectrum complexity and disturbance presence cause,
that diagnosis accuracy largely depends on expert
knowledge and experience;
Diagnosis on the basis of current spectrum depends on
harmonics amplitude valuation with characteristic
frequencies for given failure;
REFERENCES

Table 1. Theoretical harmonics distribution in stator current spectrum for SSh

90L4 engine

fs=50Hz, pb=2, s=0,0366, nw=1, k=1, Nr=26

Component
de, nd = 1
PSH, nd = 0
de, nd = 1
Frequency
652
676
700
de- dynamical eccentricity, PSH- principal slot harmonic

[1]
[2]

Cz. Kowalski Monitorowanie i diagnostyka uszkodze silnikw

indukcyjnych z wykorzystaniem sieci neuronowych Oficyna
Wydawnicza Poltechniki Wrocawskiej, Wrocaw
S. NANDI, Condition monitoring and fault diagnosis of electrical
motors, IEEE Transactions on Energy Conversion, Vol. 20, No. 4,
December 2005, str. 719-729S

On figure 2 there are presented chosen examples of stator

current spectra for SSh 90L4 inductive motor.

Figure 2. SSh 90L4 motor stator current spectrum: a) theoretical harmonics

distribution in spectrum,
b) motor without load, c) motor with nominal load and out-of-line settlement
with machine, d) motor harmonics distribution with nominal load and out-ofline settlement with machine

In table 1 there have been presented calculated frequencies,

according (4.1) equation, which may be expected in SSh 90L
4 motor stator current spectrum. The measurements were made
on the test stand equipped with Brael&Kjaer analyzer. The
results were presented on 2bd figures. From comparison of
theoretical (2a) and actual (2b, c, d) results we can conclude,
that the frequencies characterizing the eccentricity are visible
even for an engine with unknown eccentricity (non-damaged
motor), what confirms the usability of current spectral analysis
in eccentricity diagnostics.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
320

Model of a Wind Turbine with PMSG and Wind Park

Configuration
Michal Piotr Jankow
Ryerson University, Toronto, Canada
Wroclaw University of Technology,
Wroclaw, Poland, 2009

Abstract-Nowadays the number of dispersed generators (DG)

is growing rapidly. This change will greatly influence the power
system dynamics. A distribution network, where DG are
connected to the grid, cannot be considered as passive anymore.
So in future it will not be possible to use simple equivalents of
distribution networks for power system dynamic modeling as it
was before. In dynamic studies the whole power system cannot be
represented in a detailed manner because of huge system
dimensions. Therefore special techniques have to be applied for
aggregation and order-reduction of distribution networks with
DG. Future of power generation which is cleaner and more
efficient to meet the targets of the Kyoto protocol is a head of us.
Distributed energy resources systems are small-scale power
generation technologies used to provide an alternative to or an
enhancement of the traditional electric power system. Therefore,
distributed generation will play a vital part in the future of
electricity generation. However the connection of distributed
generation may pose several problems to local distribution
networks, therefore, analysis and tackling these problem areas
will be needed to address new challenges for a network design,
operation and modeling of a power system is required of utmost
importance. In this paper brief review of load and generation
models have been given.

I.

INTRODUCTION

As we strive to create a future of power generation which is

cleaner and more efficient to meet the targets of the Kyoto
protocol and reduce the dependency of our now limited
resources of fossil fuel, distributed generation will play a vital
part in the future of electricity generation.
Currently, industrial countries generate most of their
electricity in large centralized facilities, such as coal, nuclear,
hydropower or gas powered plants. These plants have excellent
economies of scale, but usually transmit electricity long
distance. Most plants are built this way due to a number of
economic, health and safety, logistical, environmental,
geographical and geological factors.
Distributed generation (DG) is another approach. It reduces
the amount of energy lost in transmitting electricity because the
electricity is generated very near where it is used, perhaps even
in the same place. This also reduces the size and number of
power lines that must be constructed. Therefore, distributed
energy resources systems are small-scale power generation
technologies used to provide an alternative to or an
enhancement of the traditional electric power system. It is the
expectation that electricity customers will wish to invest in

their own local generation system which will make use of these
DG technologies [1], [2].
However the connection of distributed generation plants may
pose several problems to local distribution networks. Analysis
and tackling these problem areas will be needed to address new
challenges for network design and operation as well as looking
for new methods of modeling of power system network is
required of utmost importance. Therefore, power control
performance of the DG unit determines its impact on the utility
grid it connects to.
The process for analysis have been automated, with the
addition of adapted existing load and generation profiles being
put in place to provide realistic load and power flows within
the network. This analysis will be used to provide an insight to
the anticipated behavior of distribution network allowing for
the suitability of alternative novel network technologies to be
assessed to wider development [2].
II. DESCRIPTION OF GENERATION MODELS
In recent years, the problem of energy crunch has become more
and more aggravating, resulting in increased exploitation and
research for new power energy resources around the world.
Natural resources in the world have depleted rapidly as
mankind venture into the new millennium. The energy
consumption is steadily increasing and the deregulation of
electricity has caused that the amount of installed production
capacity of classical high power stations cannot follow the
demand [1]. A method to fill out the gap is to make incentives
to invest in alternative energy sources such as photovoltaic
systems, biomass and hydropower resources.
III. MODEL OF WIND TURBINE WITH PMSG
As numerous WTG systems have been developed and
connected to power systems worldwide, many different wind
turbine models have been developed. Three of the most
popular wind turbine generator topologies are outlined in Fig. 1
namely, the fixed-speed design employing a squirrel-cage
induction-machine and variable-speed designs using the
doubly-fed induction generator (DFIG) and the multi-pole
synchronous machine.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
321

Each of the wind turbines in Fig. 1 comprises a prime mover

and an electrical generator/power system interface. The prime
mover consists of the wind-turbine blades, which may include
a pitch controller, the wind turbine shaft, and possibly a
gearbox. The electrical generators typically comprise squirrelcage induction machines, wound rotor induction machines, o
synchronous machines.
Traditionally, a wind turbine with PMSG is connected to the
AC grid through back-to-back full converters. Nowadays, with
the evolution of power electronics, voltage source converter
based HVDC (VSC-HVDC) has been considered as a feasible
solution to integrating wind farms to grids due to its favorable
features. The analysis in the paper is under the background of
VSC-HVDC.
A. Generator Model
The dynamic model of PMSG is derived from the two-phase
synchronous reference frame in which the q axis is 90 ahead
of the d-axis with respect to the direction of rotation. The
synchronization between the dq-rotating reference frame and
the abc-three phase frame is maintained by a phase locked loop
(PLL).
The electrical model of PMSG in the synchronous reference
frame is given in (1),(2).

where subscripts d and q refer to the physical quantities that

have been transformed into the dq-synchronous rotating
reference frame; Ra is the armature resistance; e is the
electrical rotating speed which is related to the mechanical
rotating speed of the generator as e = npg where np is the
number of pole pairs; 0 is the permanent magnetic flux. The
electric frequency is determined by fe = e/2. The frequency is
obtained by PLL. The inductances, Ld and Lq are the
summation of the inductances of the generator on the d- and qaxis and the inductance of the transformer L1, respectively. ud
and uq are, respectively, the d- and q-axis components of ug in
Fig.7. The q-axis counter electric potential eq= e0 and the daxis counter electric potential ed = 0. In the paper, we assume
Ld = Lq = L and (3), (4) can be rewritten as
did /dt = (-Ra /L)id + eiq + (1/L) ud
diq /dt = (-Ra /L)id e[id + (1/Lq) 0]+ (1/L) ud

(3)
(4)

For the complete representations of the wind turbine with

PMSG the following models are required: Aerodynamic, Pitch
Angle Control, Speed Control and Phase Locked Loop. In the
[3] all of them have been described clearly.
PMSM model could also be modeled by linearization
technique. This proposed method [4],[5], removes the
dominant quadratic nonlinearity of the model as well as higher
order terms involving the input in the system. PM synchronous
motor can be linearized using the following transformations
[5],[6]
y = x + (x)

(5)

u = (1+ (x))v + a(x)

(6)

x = Ax + Bu + f2(x)

(7)

x = [x1 x2 x3] T, u=[u1 u2] T

(8)

f2(x) = [k1 x2 x3; k2 x3 x1; k3 x2 x1]

(9)

where k1, k2, k3 are as defined as:

Fig. 1. Fixed-speed, DFIG, and multi-pole synchronous WTG

topologies.

did /dt = (-Ra /Ld)id + e(Lq/Ld) id + (1/Ld) ud

(1)

diq/dt = (-Ra /Lq)iq - e[(Lq / Ld)id + (1/Lq) 0] + (1/Lq )uq (2)

k1 = (Lds Lqs)/a31,

(10)

k2 = (-Ldsa31)/Lqs,

(11)

k3 = (LqsC12a31)/Lds

(12)

where a31 and C1 is :

a31 = (3P2LdmIf/8J), C1 = 1/Lqs

(13)

Lf is the field current equivalent to the permanent magnet. Lds,

Lqs and Ldm are the direct, quadrature and magnetizing
inductances respectively. Rs is the stator resistance. P is the

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
322

number of poles and J is the system moment of inertia. A and

B are in Brunovsky form as in:

The simplest way to represent the wind farm is to model the

entire farm as an equivalent single wind turbine. This approach
assumes that the power fluctuations from each wind turbine are
A = [0 1 0; 0 0 1; 0 0 0]
(14) all equal throughout the farm. This assumption, however, does
B = [0 0; 0 1; 1 0]
(15) not reflect reality, because the power fluctuations of a wind
farm are relatively smaller than those of a wind turbine.
Equation (7) can be linearised using the transformations (5) Another way to model the wind farm is through a detailed
and (6) (with (6) modified for the case of two inputs) where
modeling of the farm and considering factors such as the
coherence and the correlation of wind turbulence as the
(x) = 2(x) = [0; k1 x2 x3; k1 x23], (x) = [-k3x2x1; -k2x3 x1] (16) presented in. These models imply a heavy load of
mathematical modeling and sizable hardware to process them.
(m-1)(x) = (-1)m-1[BT{2(x)/x}B] m-1, m 2
(17) The model presented in this work takes into account the
aggregation effect of the wind farm using an equivalent for the
The system then reduces to y = Ay + Bv. Proof: See [5],[6].
wind added to groups of wind turbines in the farm. The model
thus conformed renders a good approximation of the behavior
of
the wind farm, from the electric system viewpoint. As an
IV. WIND FARM MODEL
advantage, the need for computational resources is reduced [7]
The mathematical model used for the wind farm behavior in
power systems is presented in this section. Typically, the
number of wind turbines in a wind farm is high. The example
of wind farm is shown in Fig. 2. In fact, a large wind farm can
feature hundreds of wind turbines. Therefore, only for wind
farm projects it is necessary to analyze in detail the entire
generating facility, with each wind turbine represented
individually.

Fig. 3. Dynamic model of an asynchronous generator.

A. Proposed Linear Dynamic Model Applied to a Wind Park

In [8] induction generator dynamic model was presented as a
Thevenin equivalent voltage source E behind the impedance Rs
+ jX as can be seen in Fig. 3. The proposed linear dynamic
model can be applied to a wind park consisting of several
asynchronous generators. In this case the variation of stator
current, for machine number i, is expressed as:
IS,I = (UI EI) /(RS,I+JXI)

(18)

where Ui is the nodal voltage defined by nodal analysis of the

circuit.
U = KE

(19)

Fig. 2. Example of wind farm model.

In studies where the objective is to verify the influence of the

wind farm on the electrical system, the model of every
individual turbine of the wind farm would need excessively
long processing times and a very robust computational
infrastructure. In such studies, the wind farm is represented by
an equivalent model from the viewpoint of the electrical
system [7].

Applying equations which are derived in detail in [20], the

result is a matrix expression such as

(20)

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
323

where xi 3 x 1 vector, Ai,i 3 x 3 matrix with parameters of

machine i, Ai,k 3 x 3 matrix that represents the relationship
between variables of machine i and k, bi 3 x 1 vector.
In [8], more general expressions for this induction generator
model are presented. Therefore steady-state situation of (20),
for sinusoidal fluctuations of Pm can be obtained by
superposition. Therefore, for Er, Em and s, the following
general expression can be written


   ,
SIN,
  ,
COS ,




21

REFERENCES
[1]

[2]
[3]
[4]

[5]

V. CONCLUSIONS
In this paper brief review of load and generation models
have been given. The whole next part of this paper was
dedicated to wind energy. The paper presents a model of the
variable speed wind turbine with PMSG that is scheduled to
connect an AC grid through VSC-HVDC. The complete model
consists of a PMSG model, a pitch-angled controlled wind
turbine model and a drive train model. The generator rotational
speed control scheme features the concept of vector control in
the
dq-synchronous
rotating
reference
frame.
MATLAB/Simulink was implemented to build the dynamic
model of the wind turbine with PMSG [3]. PMSM model could
also be modeled by linearization technique. This proposed
method [4],[5], removes the dominant quadratic nonlinearity of
the model as well as higher order terms involving the input in
the system. Next, a linear dynamic model for an asynchronous
wind turbine and for the wind farm with mechanical sinusoidal
fluctuation was developed. This proposed linear model
presents good accuracy compared to the dynamic model. The
proposed model has the following advantages: 1) it has low
complexity; 2) it permits modal analysis, direct solution of
differential equations, and easy implementation in several
conventional tools for power system analysis (load flow
analysis and power system stability) [8].
Future of power generation which is cleaner and more
efficient to meet the targets of the Kyoto protocol is a head of
us. Therefore, it is needed to address new challenges for a
network design, operation and modeling of a power system is
required of utmost importance.

[6]
[7]
[8]

M. Dai, M. N. Marwali, J. W. Jung, and A. Keyhani, Power flow control

of a single distributed generation unit with nonlinear local load, in Proc.
IEEE PES Power Syst. Conf. Expo (PSCE04), New York, NY, Oct.
2004, vol. 1, pp. 398403.
S. J. Haig, R. M. Tummity, G. M. Burt, J. R. MacDonald, Analysing the
technology needs of future distribution networks, UPEC 06, 41st
International Proceeding, Vol.1 6-8Sept. 2006, pp. 217-221
[Link], L. Gengyin, Z. Ming, Z. Chengyong Modeling of the Wind
Turbine with a Pernament Magniet Synchronous Generator for
Integration, IEEE PES General Meeting ,24-28 June 2007.
[Link], [Link], R. Devanathan,"A New Linearisation
Technique for Pernament Magnet Synchronous Motor Model, Pwer
System Technology, IEEE Power India Conference 2008, POWERCON
2008, 12-15 Oct. 2008.
[Link], Aruna Rajan, R. Devanathan,"Complete Quadratic
Linearisation of PM Synchronous Motor Model ", NPEC 2005, pp 49-52.
[Link], Aruna Rajan, [Link]," Linearisation of PM
Synchronous Motor Model", Technical Report, Department of EEE,
Hindustan College of Engineering, Tamil Nadu 2005.
Y. Coughlan, P. Smith, A. Mullane, M. OMalley, "Wind Turbine
Modelling for Power System Stability Analysis A System Operator
Perspective", vol. 22, pp.929-936, Aug. 2007.
J. CIdras, A. Feijoo, "A linear dynamic model for asynchronous wind
turbine with mechanical fluctuations", IEEE Transactions on Power
Systems, vol. 17, pp.681-687, Aug. 2002.

BIOGRAPHY
Michal Piotr Jankow was born in Wroclaw, Poland, in
1985. He is currently studying for a [Link]. degree in
Control in Electrical Power Engineering at Wroclaw
University of Technology, Wroclaw, Poland and for a
MEng. degree at Ryerson University, Toronto, Canada.
His present interests are wind energy, wind turbines,
offshore and onshore wind parks. His also the finalist of
the national competition by LOREA'L in the topic:
"Applying renewable energy sources in industry".

ACKNOWLEDGMENT
I would like to thank all the authors, and those advancing in
power energy systems and distribution technology worldwide.
This special issue would not be possible without your
achievements also I wish to thank the IEEE and to all authors
for their content-related support and materials who provided all
the research in load and generators modeling so this paper
could be done in pulling all the papers together which you can
find in references.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
324

Load Modeling For Power System Analysis

Jakub Kepka
Wroclaw University of Technology
AbstractThis paper covers overview of the load models for
distribution systems. The basic concepts for static, dynamic
and combined models are introduced. The following discussion
shows importance of accurate system parameters modeling for
analysing power system stability. The accurate load
representation is not easy issue regarding to its changeable
nature and variety. Basic load classification is described.

represents variation of the load with the frequency. The

fundamental starting point for the load modeling is at the
distribution level, thus, the applications outside of power
system can be as follows:[1]
Static applications this model considers only voltage
depended characteristic[1]:

I. INTRODUCTION

o
o
o

ower system stability and reliability have been very

important issues over the years. It is hard to predict the
system behaviour in case of outage, voltage or frequency
variation or other factors that can affect system stability
without proper information, or system behaviour data. The
powers system are mostly nonlinear and operates in
constantly changing conditions, so it is hard to obtain
precise information and data about i.e. generators or loads
during disturbance[2]. This required to either assume or
measure certain quantities like real and reactive power, to
predict system response to abnormal conditions. Very
important factor is to not interrupt power flow to the
residential, industrial or any other end users. That is why
certain measurements need to be made with specialized,
expensive equipment to allow get the data under operating
system. This type of data collecting involves a continuous
monitoring process and requires data to be processed after
the measurements are done.[3]. A lot of studies have been
carried out so far to provide as accurate model of load as
possible, and still new techniques are being developed
through big electric concerns. There is a lot of computer
software existing that allows for load modeling and system
behaviour simulation, however, as mentioned above, the
more accurate data typed in to the program, the more
accurate model and the simulation [Link] has become
clear that assumptions regarding the load model can impact
predicted system performance as significantly as the models
chosen for excitation systems and synchronous machines,
which have received greater industry attention[5]
Analysis of power system dynamic performance requires
the use of computational models representing the nonlinear
differentialalgebraic equations of the various system
components. While scale models or analog models are
sometimes used for this purpose, most power system
dynamic analysis is performed[2].

Power Flow (PF)

Harmonic Power Flow (HPF)
Transmission System (TPF)
Transmission Power Flow
Voltage Stability

Dynamic applications considers both voltage-dependent

and frequency dependent characteristics[1]:

Transient Stability (TS)

Dynamic Stability (DS)
Operator training simulators (OTS)

As one already mentioned, the loads are quite difficult to be

modeled in terms of their variety and change in load
depending on the time of the day or season. It is necessary to
classified them according to particular area, composition of
each one as well as regarding to load characteristic. The
loads can be also combined in groups and exhibited as:
industrial, residential, commercial and agricultural[3]. The
industrial loads includes mainly induction motors, up to
95%, the residential loads includes most of domestic
appliances, i.e. refrigerators, washing machines etc. as well
as heating and air conditioning units. The commercial load
is refers to air condition units and discharge lighting in
particular, whereas agricultural loads corresponds to
induction motors, i.e. as a prime movers for pumps.[3]
Having presented load classification and general load
requirements, it is possible now, to carry out the theoretical
analysis to express the system load as a mathematical
models. In order to do this, one needs to use some
mathematical expressions to describe the load as accurate as
possible, what is described in the next section of this paper.
III. LOAD MODELING CONCEPTS

Generally, one can distinguish tow basic load modeling

concepts: component based and measurement based models.
Component based model is built on the base of the
II. LOAD TYPES AND CLASSIFICATION
information on each elements that consist of the load being
Before load modeling process will get started, it is
simulated. Each load component type is tested and to
required to understand its applications that are divided into
determine the relationship between real and reactive power
two categories: static, called snap-shot with respect to requirements versus applied voltage and frequency.[1],
time as well as dynamic time varying. Since the static then the load model is developed mainly in either
model based
exponential or polynomial form basing on obtained
[Link] range of validity of each model is directly related
to the range over which the component was tested[1].
on the steady state method of the network representation in Usually, the load model is expressed on the p.u values for
the power flow networks[1], these models express loads as convenience. By combining appropriate load model types
a voltage magnitude
function. The dynamic model, in turn, one can
approximate
the Poland,
composite
load. 2009
Mentioned load
8. EEEIC International Conference on Environment and Electrical
Engineering,
Karpacz,
May 10-13,
325

model types are obtained basing on the load survey

information.[1] The composition of the load depends upon
the time of the day, moth, or season as well as atmospheric
conditions. One can notice, that the country with fine
weather revel higher demand for air conditioning, while in
the cold climate countries it is just opposite, there is a big
demand for electricity heating. One can point out, that these
demand can also vary seasonally, in the summer/winter
periods, demand for heating/cooling can either increase or
decrease.[3] Having introduced the model types and its
variation depending on certain factors, one can propose
some mathematical load model representation as a set of
equations in order to give details of the relationship between
the input and output system parameters. Considering load
modeling, mathematical representations is related to voltage
and/or frequency measured at the buses as well as real and
reactive power consumed by the loads.[3]
IV. LOAD MODELING
For dynamic performance analysis, the transient and
steady-state variation of the load P and Q with changes in
bus voltage and frequency must be modeled [5]. For power
system analysis load can be thought as real and reactive
power launched to lower voltage distribution system at
buses represented as network model. Among lots of devices
and appliances being connected to the system and
considered as a load, one should include intervening
distribution systems feeders, shunt capacitors, transformers
etc. as well as voltage controlling devices or voltage
regulators. The simplest load model is static model. Next
chapter introduces static models as well as dynamic ones
that are being used for i.e. system stability studies.
V. DELTA AND WYE CONNECTED LOADS
First load models to be described are Wye and Delta
connected loads. Each of these loads can be represented as
being connected phase-to-neutral or phase-to-phase in a four
wire Wye system or phase-to-phase in a three wire delta
system[4]. All of model above can be three, two, or single
phase and modeled as: constant real and reactive power,
constant current, constant impedance as well as any random
combination of the above. This model is one of the basic
static load models.

Fig.2 Delta connected load[1]

Both load models depicted in the figures above require

the load component current coming into the load to be
determined.[r4]. Assumption is made, that the loads have an
initial complex power
||,, = ,, + ,, (1)
, where indices a,b,c denotes that formula is expressed in the
same way for each single phase of Wye connected load, this
notation will be kept in further formulas describing model
above
A. Constant Real and Reactive power loads for Y
connection
In this model the line to neutral voltage will change during
each iteration until convergence is achieved[4]
,, =

,,

,,

||,,

,, ,, = (2)

,,

B. Constant Impedance Load for Y connected loads

Constant impedance is first determined from the specified
complex power and line-to-neutral voltage for each phase.
,, =

,,
,,

,,
|,, |

,, = |,,| ,, 3

Having constant load impedance, it is possible to express

load currents as the function of those impedance. This
current for each phase is given by equation (4):
,, =

,,
,,

||,, ,, (4)

|,, |
|,, |

,, ,, =

In this model the line to neutral voltage will change during

each iteration until convergence is achieved[4]
C. Constant Real and Reactive power loads for Y
connection
Fig.1 Wye connected load[1]

,, = |,, ,, ,,

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
326

For this model, current magnitudes are obtained according

to the equation (2). These magnitudes do not change until
the voltage angle changes during each iteration[4],
D. Delta connected load model
Similarly to the Y connected load, the delta connected load
model can be made. The equations describing this model are
the same, except that voltages and currents in this models
are taken as phase-to-phase instead of phase-to-neutral, so
delta load model is described by following equations:
Constant real and reactive power:
||
| | = 6
= =

Constant Impedance Load

=

|| 2

|| 2
||

|| = 7

Constant Current Load

,, =

||
||

( || = 8)

, where index x denotes line-to-line values ab, bc, and ca

respectively.
Models presented above can be modeled as combination of
above constant Z, I, and P and can be assigned as a
percentage of total load shown above. For the Wye
connection, it is simply the sum of the three components Ia,
Ib and Ic, while for all delta connected loads the line current
is determined by:

Fig. 4 Delta connected capacitor bank[1]

The capacitor banks are modeled as constant susceptances

connected as above. The susceptance can be found as:

10
2 1000
.

.=
11
2
.

12

.=
13
___
=

Having computed susceptance one can find out the line

currents given by:
= 14
= 15
= 16
For delta connection currents flowing in capacitors are given
as below:
= 17
= 18
= 19

Considering Wye and Delta connected loads one can

mention of the Shunt Capacitor models that are very often
used for voltage regulation in a distribution system as well
as in order to provide reactive power support.

The line currents feeding the delta connected capacitor bank

are as follows:

= 1

0
1
1

1
0
1

20

VI. POLYNOMIAL AND EXPONENTIAL LOAD MODELS

There are two mathematical forms that are commonly used
in modeling studies: polynomial and exponential load
models. Polynomial model is a static model describing
relationship of the power and voltage magnitudes as a
polynomial equation and can be expressed as[6]:

= 0 1 + 2 + 3 21
0

Fig. 3 Wye connected capacitor bank[1]

= 0 4

+ 5 + 6 22
0

8. EEEIC International Conference on Environment and Electrical

Poland,
May 10-13,
,while Engineering,
exponentialKarpacz,
model has
the form
of: 2009
327

= 0

= 0
0

23

24

The parameters of this model are exponents, and the power

factor of the load[6]. Setting these parameters to 0,1, or two
the load model becomes constant current, power, or
impedance, like in case of the Wye and Delta connected
loads shown in previous section.[6]
A. PSS/E Static Model[12]
This static model introduces the active and reactive power
that are both frequency and voltage dependent and given by
the following formulas:
= 0 1

+ 2 + 3 1
+ 7

(25)

4
5
= 0 4 + 5
0
0
6
+ 6 +8
0
(26)

[5]
M. Sedighizadeh, and A. Rezazadeh, Load
Modeling for Power Flow and Transient Stability Computer
Studies at BAKHTAR Network
[6] IEEE Task Force on Load Representation for Dynamic
Performance,
Load
representation
for
dynamic
performance analysis, IEEE Transactions on Power
Systems, Vol. 8, No. 2, May 1993
[7]
Byoung-Kon Choi, Member, IEEE, Hsiao-Dong
Chiang, Fellow, IEEE, Yinhong Li, Yung-Tien Chen, DerHua Huang, and Mark G. Lauby, Development of
Composite Load Models of Power Systems using On-line
Measurement Data
[8] Jingchao Zhang Anhe Yan Zhuoya Chen and Kun Gao
Dynamic Synthesis Load Modeling Approach Based on
Load Survey and Load Curves Analysis
[9]
Pablo Ledesma, Member, IEEE, and Julio Usaola,
Member, IEEE. Doubly Fed Induction Generator Model for
Transient Stability Analysis, IEEE TRANSACTIONS ON
ENERGY CONVERSION, VOL. 20, NO. 2, JUNE 2005

[10] Zhijun Li and Lijuan Cui Department of Electrical

Engineering and Automation, Hebei University of
Technology, China. The Building and Analyzing of the
Fifth-Order Model of the Synchronous Generator in StandAlone Infinite System
[11] Prabha Kundur, Neal J. Balu, Mark G. Lauby. Power
System Stability and Control, McGraw-Hill Professional,
1994
[12] Y. Li, H.-D. Chiang, B.-K. Choi, Y.-T. Chen, D.-H.
Huang and M.G. Lauby. Representative static load models
for transient
stability analysis: development and examination, McGrawHill Professional, 1994

,where a1- a8 and n1- n8 parameters to be estimated when

V=V0

VII. CONCLUSIONS
The both static, and dynamic model of loads were described
in this work as well as combined ones. Load modeling is a
challenge for today`s engineers. It`s variety and not constant
nature makes them hard to simulate. There have been a lot
of studies on load modeling in order to represent it with
maximum accuracy. There are many simulation software
that allows to simulate power system loads basing. i.e. on
polynomial and exponential load model representation.
VIII. REFERENCES
[1] Prabha Kundur, Power System Stability and Control,
2007
[2] Prabha Kundur, Electric Power Generation,
Transmission, and Distribution
[3] Ins Romero Navarro, Dynamic Load Models
for Power Systems, Estimation of Time-Varying Parameters
During Normal Operation
[4] William H. Kersting, Electric Power Generation,
Transmission, and Distribution
8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
328

Loads and Synchronous Generator Modeling

Robert Pliszczak
Wroclaw University of Technology
Abstract-- This paper survey of different accurate models of
load and synchronous generator model. I would like to described
one out of two basic types of load modeling. One is componentbased load modeling which contain ZIP (polynomial) model,
exponential model, combination of this two called exponentialZIP model and ETMSP & EPRI model as an example of
combination of fundamental models. Second one is measurementbased load modeling, however, this is not on my consideration.
The second part survey consists synchronous generator modeling
by electrical circuits. Moreover, there are dc generators, but this
model is not covered in this paper.

I.

INTRODUCTION

here are two general issues to consider modeling of power

systems. One of them is loads modeling as well as
generators modeling. Loads and generators modeling are very
important in determining the stability of power systems.
Moreover there are valuable tools in the stability studies.
Load modeling application can be distinguished into static
and dynamic models. First of all I would like to shortly
describe static load model. The static models were discovered
to investigate steady-state power systems. In this models
occurs constant voltage magnitude, but on the other hand there
is variation of load frequency. Lastly the static load models
feature is voltage-dependent characteristic. Second of all it is
mentioned in the technical literature about dynamic load
modeling. This load model is in time-domain and contains
loads electromechanical behavior. Phasors method is the load
steady-state representation. The dynamic load model is based
on time variation frequency. Moreover, the dynamic load
model is describe by dynamic voltage and frequency
characteristics.[1]
The second part of this paper is to survey generators
modeling. The Generators can be classified as a electrical
motors. There are dc generators, synchronous generators and
inductor generators. The dc generators are rarely used in power
systems. The most popular and common use in the industrial
applications is synchronous generator. There have been
developed big number of synchronous generators. The
generators are the heart of power systems. With the
increasing cost of detailed prototyping of electrical machine, it
is becoming necessary to replace or supplement it with
mathematical methods and computer simulation.[2] However,
the inductor generators are used in power systems, but mostly
in wind turbines. The number of wind turbines connected to the
grid increase every year. Papers [8], [9], [10] are contained
more information`s about double fed induction generators
modeling.

II. LOAD MODELING

Load modeling is a very important issue in power system
analysis and control. Accurate load models have significant
impact on the operations flexibility and exploitation costs in
power systems. Exploitation costs can be reduced because of
there is possible to test new solution and change power system
configuration without any disturbances of normal power
system work. The operations flexibility is also very important
regarding to beater fit power system modeling into real power
system. That is why both power industry and academia are
looking for now, and more accurate loads representation. We
can distinguish two approaches for load models developing:
component-based and measurement-based. The componentbased approach builds a load model using detailed information
about all the individual components at each load bus. The
measurement-based approach involves placing measurement
systems at the load buses for which load models are to be
developed.[3] The measurement-based approach gives us
actual load behaviors at any given moment. In the power
industry is also used on-line data recording system. This tool
gives important information about actual load behaviors during
system disturbances. Dynamic load models regarding to
possible disturbances response are highly recommended for a
lot of power system models representations, moreover this
models are commonly used in power industry for transient
stability analysis. On the other hand, it increases of system
representation for simulation. [3]
Loads modeling has always been a difficult issue because:

many different types of load are connected to the
power system at any period of time,

level of complication and quantity of data is very
high at any given moment,

it is hard to predict loads response (behavior), [4]
We can distinguish three main load models:
 ZIP (polynomial) model,
 exponential model and
 exponential ZIP model - combination of this two
above [1]
A. Polynomial-ZIP Model
One of the oldest load representation is polynomial model.
It is essentially combination of constant impedance (Z),
constant current (I) and constant power (P), that is why it is
commonly called ZIP model. This model is not accurate
representation of the system, because of high number of
inductor motor loads in the system. Nowadays there are more
complex and accurate load models which capture both the
static and dynamic load, such as induction motors.[4]

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
329

C. Exponential-ZIP Model
This is the combination of both exponential and ZIP
models which are shown above. Diagram can be the same as in
previous model.



 1    !     " 





#$
 
 !   

% 1   


#$
  

1   

Fig.1. The structure of ZIP load model. [5]




 
  
  
   1   





    
  
   1   



 and  are the constant impedance load parameters,

 and  are the constant current load parameters,
 and  are the constant power load parameters,
 and  are the frequency sensitivity parameters,
and       1,       1 when   . [2]
B. Exponential Model



 1    !     " 




#$
% 1   
 
 !   



#$
  

1   


 and  are the constant current parts of the total load,

! and ! are the constant power parts of the total load,
 ,  , &'1,  ,  , &'2,  ,  , &'1,  ,  ,
&'2 are the parameters of exponential part of the total load.[3]
D. ETMSP & EPRI Model
Scheme is shown in Fig.2. However, there are some more
models discoverd by the companies. The companies which
have made distribution system computers programs (software).
They have used this three types of load modeling and modify
them. EPRIs LOADSYN and ETMSP packages are one of the
most commonly used software for dynamic studies. This two
packages based on load models presented above with little
changes. Below the result of this changes is shown:
, ,1
1     .1 /
+ 0 

2

 
 *
+ 




Fig.2. The structure of any static load models.[6]

This scheme can be used for any static load modeling. So, it fit
for every presented in this paper load model.


 
 

1   



   

1   


 -voltage-dependent parameter of real power,

 -voltage-dependent parameter of reactive power,
 and  are the frequency sensitivity parameters.[3]
Comparing of ZIP and exponential load modeling the second
one is more accurate.

, 
 *+ 

1   


,1
  / +


1   2




where,

+ is the frequency dependent fraction of real load,
+ is the reactive load coefficient of uncompensated reactive
load to real power load,
$ and $ are the voltage exponents for frequencydependent and frequency-independent real power load,
$ and $ the voltage exponents for the uncompensated
and compensated reactive power load,
 and  are the frequency sensitivity coefficients for
real and uncompensated reactive power load
 is the frequency sensitivity coefficient for reactive
compensation [3].

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
330

We have showed this load model as a example of one

combination of this fundamental models.
III. GENERATORS MODELING
Generators convert mechanical energy to electrical energy.
There are three possible types of generators: dc generators,
induction generators and synchronous generators. The first one
are useless because all the power systems around the world are
AC. So there are two main types of generators: one of them is
synchronous generator and second is induction generator. The
inductor generators mostly occur in the wind mills. The
synchronous generators are common use in most types of
power plants.
First of all the synchronous machines work with constant
speed what means constant frequency. The speed of the
synchronous generator linearly depends on the frequency and
numbers of poles. The constant frequency is very good feature
because one of the power systems requirements is to keep
frequency in one specified level. Another advantage comes
from reactive power compensation in the grid. The
synchronous generators produce active power and reactive
power. The last important feature is high efficiency of
synchronous generators.
The inductor generators are less popular then the
synchronous generators. However, they have some advantages.
One of them is that inductor generators are simplified in
construction. Furthermore, they are lighter and cheaper than
any others. However the frequency cannot be adjusted as
simple as in the synchronous generator and the reactive power
cannot be compensated. When one would like to use inductor
generator it should also have special control system. The speed
controllers are going to be cheaper and cheaper and inductor
generators are commonly used in the wind mills.
I would like to present synchronous generator modeling by
electrical circuits. There are more synchronous generators
models in papers [7],[11], [12], [13].

A. Synchronous Generators Modeling by Electrical

Circuit
From figure above the machine equations in three axis
framework is shown below:
7
'+3!  /45 6+3!  +3!
78
7
'  4 6  
78
7
0  4; 6;  ;
78
7
0  4< 6<  <
78

6; and 6< are the direct and transverse dumpers` currents,

6 is the exciter current,
; and < are the direct and transverse dumpers` total flux,
+3! is the stator total flux,
 is the main field total flux,
The Park`s matrix can be written as:
2 BCD.=> 0

.=> 0  ? A
3
/D6&.=> 0

2E

3
2E
/D6& => /

3
BCD => /

2E

3 F
2E
/D6& => 

3
BCD => 

'+
'K

.=> 0 '+3! 
.=> 0 H'3 I  J ' L

'!

We can rewrite first equations as:

7
'K  /45 6K  K / M> 
78
7
'  /45 6K   / M> K
78
7
'  4 6  
78
7
0  4; 6;  ;
78
7
0  4< 6<  <
78

Then the following relationships can be written considering the

generator convention for stator

Fig.3. Synchronous generator winding with dampers.

D and Q represent d-axis and q-axis dampers. The three

synchronous generator modeling methods presented below are
based on this scheme.[2]

K  +K  R5K  RK 
Q
O S+K /6K  6;  6  / SR5K 6K  SRK 6 / 6K 
  +  R5 
O
O
 S+ /6  6<  / SR5 6
O
  +K  R  RK 
T
P  S+K /6K  6;  6   SR 6  SRK 6 / 6K 
O
;  +K  R; 
O
 S+K /6K  6;  6   SR; 6;
O
<  +  R< 
O
 S+ /6  6<  / SR< 6<
N

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
331

From all this equations we can draw the synchronous machine

electrical scheme as follows:

Fig.4. d-axis and q-axis electrical equivalent circuits. [2]

R is the rotor leakage flux,

R; and R< are direct and transverse dampers leakage flux,
R5K and R5 are direct and transverse stator dampers
leakage flux,
RK is the linkage flux between main field and stator d-axis,
+K and + are direct and transverse main flux,
SR5K and SR5 are direct and transverse stator leakage
inductances,
SR is the rotor main field leakage inductance,
S+K and S+ are the direct and transverse stator main
inductances,
SR; and SR< are dampers leakage inductances,
SRK is the linkage inductance between the rotor main field and
the stator d-axis,
In these scheme the linkage inductance between d-axis stator
and direct damper is neglected. In the Fig.4. the blocks "dload" and "q-load" represent the equivalent loads in Park`s
framework of the real power factor load. This last is starconnected and an inverse Park`s transformation is needed to
elaborate "d-load" and "q-load". This is useful because all
equations involved in the simulation are given in Park`s
framework [2]
IV. CONCLUSIONS
In this paper, different accurate models of load and
generation were examined. As a load models component-based
load modeling were shown which contain ZIP (polynomial)
model, exponential model, exponential-ZIP model and
ETMSP & EPRI model as a example of combination of
fundamental models. All models examined in this paper are
static load models. The static load models can be used for
approximation of the dynamic behavior of the power system.
Moreover static nonlinear load models can be used for transient
stability analysis. However this model is not appropriate for
capturing true load response and for industrial applications

measurement-based models are required. ZIP model gives

only approximate results of real power behavior. Exponential
model is more accurate and has only two parameters to be
estimated. Exponential-ZIP model is also very promising but
it is a little bit more complex what can give unwanted
confusion [3]. With continued expansion of power grid, it
becomes more and more important that the model used in
simulation programs consist with the actual characteristics of
electric apparatus. Up to now, the model of generators,
transformers and power lines have been established perfectly,
as for the model of load, due to its randomness, time-varying
and distribution property, it could barely describe the whole
characteristics exactly by one or more groups of equations.
Accurate load modeling has become one of the most difficult
problems in power system research fields.[5]
Second part of the survey consists synchronous generator
model by electrical circuit. This model consists the global
electrical scheme of the generator by flux equations. Paper [2]
contains tests of the three synchronous generator models which
can be distinguished into: synchronous generator modeling by
electrical circuits, by state equations using an inside load and
by state equations using an outside load.
V. REFERENCES
[1]
[2]

[3]

[4]

[5]

[6]
[7]

[8]

[9]
[10]

[11]

[12]

[13]

Prabha Kundur, Power System Stability and Control, 2007

Emile MOUNI, Slim TNANI and Grard CHAMPENOIS,
Comparative study of three modeling methods of synchronous
generator., IEEE, 2006
Y. Li, H.-D. Chiang, B.-K. Choi, Y.-T. Chen, D.-H. Huang and
M.G. Lauby, Representative static load models for transient
stability analysis: development and examination, IET Gener.
Transm. Distrib., 2007, 1, (3), pp. 422431
A. Maitra, A. Gaikwad, P. Pourbeik, D. Brooks, Load Model
Parameter Derivation Using an Automated Algorithm and
Measured Data, IEEE, 2008
Jingchao Zhang Anhe Yan Zhuoya Chen and Kun Gao, Dynamic
Synthesis Load Modeling Approach Based on Load Survey and
Load Curves Analysis, IEEE, DRPT2008 6-9 April 2008 Nanjing
China
Dmitry Kosterev, Anatoliy Meklin, Load Modeling in WECC,
IEEE, 2006
Emile Mouni, Slim Tnani and Grard Champenois, Synchronous
generator output voltage control via a generalized predictive R S T
controller., IEEE, 2008
E. Spahic', J. Morren', G. Balzer and G. Michalke, Mathematical
Model of the Double Fed Induction Generator for Wind Turbines
and its Control Quality, IEEE, POWERENG 2007, April 12-14,
2007, Setubal, Portugal
Srinivas R. Chellapilla, Badml H. Chowdhury, A Dynamic model
of Induction Generators for Wind Power Studies, IEEE, 2003
Jos Cidrs and Andrs Elas Feijo, A Linear Dynamic Model for
Asynchronous Wind Turbines With Mechanical Fluctuations,
IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 17, NO.
3, AUGUST 2002
J. R. Marti and K. W. Louie, A Phase-Domain Synchronous
Generator Model Including Saturation Effects, IEEE Transactions
on Power Systems, Vol. 12, No. 1, February 1997
E. Lu, G. Bronner, A. Ilk, C. Neumeyer, S. Ramakrishnan,
MODEL for TFTR Motor-Generator (MG), IEEE, Princeton,
New Jersey, 1995
Zhijun Li and Lijuan Cui, The Building and Analyzing of the
Fifth-Order Model of the Synchronous Generator in Stand-Alone
Infinite System, IEEE, 2005

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
332

Renewable energy from the oceans

Wave power, tidal power and ocean streams
Lars Bork, Bastian Garnitz, Thomas Krisch, Carsten Schwartze
BTU Cottbus

I. INTRODUCTION
When thinking about renewable energies, wind, solar and
hydro energy typically come to mind.
To tackle climate change and all the challenges imposed by the
need to find alternative and reliable energy sources, there is
one major resource that has remained untapped until now:
marine energy.
The potential of wave energy has been recognized for long,
and mostly associated with a destructive nature. The following
paper will give the reader a short overview about using wave
and tidal energy as a renewable energy source.
Wave energy is a concentrated form of solar energy: the sun
produces temperature differences across the globe, causing
winds that blow over the ocean surface. These cause ripples,
which grow into swells. Such waves can then travel thousands
of miles with virtually no loss of energy.
The total wave energy resource is as large as world electricity
consumption, approximately 2 TW. About 10-25 % of this
resource is economically usable. [1]
Tidal power occurs because gravities of the moon and sun
combined with the rotation of the earth cause periodic changes
of the oceans. These changes are called tidal range. The
considerable level difference between high tide and ebb at
many coasts of the earth generates an enormous energetic
potential, which can be opened and used by the mankind.
Furthermore, it is possible to use the energy of ocean streams
which are caused by the complex interactions between warm
and cold layers of water.
The global theoretic potential of tidal energy is approximated
at 3000 GW, but about 100 GW are realistically recoverable.
[2], [3]
II. UTILIZATION OF THE TIDAL RANGE TIDAL POWER PLANTS
Tidal power is the result of the gravitational interaction between the moon and the earth. This gravitational force, combined with the rotation of the earth, produces a twice-daily rise
and fall of the sea level, called ebb and flood.
The period of one tide is for example 12.4 hours at the France
coast. The difference in water levels is used to convert the
potential energy into kinetic energy, by passing through the
turbines. [4], [5]

Tidal power is predictable and reliable, but it can be used

economically first at a tidal range of 3-5 meters.
At a tidal power plant the water fills during the high tide a
basin that is separated from the sea by a barrage.
During the following low tide the accumulated water is used to
drive the turbines of the power station. The turbines can work
also in the time of the filling process. Tidal barrages are built
across a suitable estuary or bay. [6]
There are several operational modes: ebb and flood generation and a two-way-generation.
At the flood generation (see at Fig. 1) the sea level is higher
than the basin level, so the water passes the barrage and fills
the basin. The barrage then is closed to dam and to create a
hydrostatic head. At the ebb generation (see at Fig. 2) the water
will pass the barrage from the basin to the sea when the sea
level has fallen under the basin level.
A single basin system can deliver electrical energy only half a
day. An option is the double-basin system. One basin will
operate in flood generation and the other in ebb generation.
Both basins are connected and it is possible to use this
construction as pump storage power plant. The main basin will
be filled at the high tide and pumped the water to the second
basin. During the ebb, the second basin will generate the
electricity. [4], [5], [7]
There are different types of turbines that are available for
use in a tidal barrage: bulb, rim and tubular turbines.
A bulb turbine is one in which water flows around the turbine.
If maintenance is required then the water must be stopped
which causes a problem and is time consuming with possible
loss of generation.
When rim turbines are used, the generator is mounted radially
around the rim and only the runner is in the flow. This turbine
is more efficient because the water flow is not so constricted.
The tubular configuration sets the runner at an angle so that a
long tubular shaft can take rotational power out to an external
generator. [6]
The idea of tidal power plants is known since the end of the
19th century. In 1913 the first tidal power plant was built in
Husum / Germany. At present there are only a few tidal power
plants in operation. The largest power plant is situated at the
river Rance / France near St. Malo with a maximum power of
240 MW. [5], [7]

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
333

In Europe approximately 100 locations exist, where OCPPs are

going to be built. The whole potential is estimated on 70 TWh
per year. Thereby 2-3% of the European current drain could be
satisfied [10].

Figure 1: flood generation with a bulb turbine

Source: [Link]; date: 25 February 2009

Figure 2: Ebb generating system with a bulb turbine

Source: [Link]; date: 25 February 2009

III. UTILIZATION OF NATURAL ENERGY FLOWS

OCEAN CURRENT POWER PLANTS
An Ocean Current Power Plant (OCPP) works similar to the
functionality of wind turbines. Both types of power plants
make use of the kinetic energy of the streaming medium, which
they are surrounded by.
Advantageously for sea turbines a lower flow speed is needed
[8]. Compared to the air current it is much easier to predict
marine current due to less important random effects. The
streaming is whether driven by the tides or by oceanic
circulations. Therefore prognoses for generated electricity are
more reliable as well [9].
The submarine turbines are anchored at the bottom of the
ocean and the outlet of electricity happens via undersea cable.
To achieve an ideal economic usage, the water depth must be
20 35 meters and the flow velocity must reach 1.5 2.5
meters per second [10]. Due to the water density, which is a
thousand times higher than air, ocean current has a much
higher power density. Compared to the face of wind turbine,
submarine rotors realize a performance that is eight times
higher. The prototype of Seaflow, an OCPP, reaches a
performance of 350 kW at 20 cycles per minute. Prototypes
using a twin rotor already reach a performance of 2.1 MW (see
Figure 3, right) [10].
The impact of saline environment is problematical due to the
process of decomposition that affects on the construction
materials of the machines. The construction of OCPPs has
effects on the speed of the ocean current and with this on
maritime ecological system as well. The locations of OCPPs
influence the barge traffic by getting into the ships routes [8].

Figure 3
left: prototype Seaflow at the west coast of Britain (photo: ISET)
right: planed OCPP (graphic: Marine Current Turbines)
Source: Quaschning, V. (2007): Regenerative Energiesysteme Technologie,
Berechnung, Simulation, 5. Auflage, Mnchen: Hanser-Verlag 2007

IV. UTILIZATION OF THE CONTINUOUS UNDULATION WAVE POWER PLANTS

A wave carries both kinetic and potential energy. The total
energy of a wave depends roughly on two factors: its height (H)
and its period (T). The power carried by the wave is proportional to H2 and to T, and is usually given in Watt per meter of
incident wave front. The energy content of a wave is
determinable by application of the linear wave theory. [11] The
total area of the seas amounts to 360.8 million km. If half of
the body of water is raised only around 0.5 m, then 0.6 EJ are
stored in the form of potential energy. [12] To use the wave
energy however only areas of offshore regions with low water
depths are possible. Thus less than 1% of the demand of
electricity could be covered by German waters. [12] The
energy of the sea waves can be used differently for power
production. Attention should be paid to the capacities of a
wave which are not always constant. That means an energy
converter is needed that works with a high efficiency at all
ranges of performance. In addition with to the construction of
the plant you have to consider the largest wave which can be
expected to accident avoidance. Furthermore the used material
must be adapted on the rough maritime climate and the
aggressiveness of sea water. [11]
At this point of time three considerable operational principles
for wave power stations are developed:
- OWC system (oscillating water column)
- Floating system (Pelamis technology)
- TapChan

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
334

A. Oscillating water column

With OWC systems a chamber with air inclusion is created.
A wave enters from the bottom in the closed chamber.
The air of the chamber is compressed and drives a turbine and
a generator. With the reverse flow of the wave, air flows again
through the turbine back into the chamber. [11], [12] For OWC
a Wells turbine is used which rotates in the same direction
regardless of the air flow, thus generating irrespective of upward or downward movement of the water column. [3] The
capacity range of these plants extends over 500 kW up to
several MW capacities. [11]

C. Tap Chan
Tap Chan is an abbreviation for Tapered Channel. The
waves in coastal area taper to an artificial ramp. The water
which has a higher potential energy is collected in an upper
reservoir. The artificial head is used to operate turbines, for
example Kaplan turbines. [11] A huge disadvantage of this
installation type is the high place requirement in the coastal
area [12].
If Sea-parks were build with a lot of such plants then approx.
30% of European power requirement could be covered with
wave energy [11].

Figure 4: Schematic of an Oscillating water column,

Source: [Link]/pix/[Link]
date: 25 February 2009

B. Floating System
Floating systems use the potential energy of a wave. These
systems consist of a tubular with an air filled lifting body. At
the hinged joints from the lifting body, hydraulic pistons are
fixed. Due to the undulation the hydraulic pistons are pressing
an operating liquid under high pressure into a compensating
reservoir. Thus a hydraulic generator is powered [11], [12].
The stationary part of the plant is deep-seated at the sea bed
[13]. The capacity of a generator amounts to approx. 750 KW.
By linkage of several generators a total output about 30 MW
could be obtained [12].

Figure 5: Floating system

Source: [Link]/azienda_en/ricerca_sviluppo/dossier_rs/img/[Link],
date: 23 March 2009

Figure 6: Schematic of a Tapered Channel

Source: [Link]
date: 25 February 2009

V. OUTLOOK
Presently there are only a few power plants that are operated
commercial, the others are for research. Therefore it is difficult
to say how much it will cost, for example the investment for
each kW.
There is still a lot to be done. As already mentioned the
potential for renewable energy from the oceans is still there.
Altogether, although tidal energy is a reliable source, however
the technically usable potential of the tidal energy appears
globally seen too low, that it could make a substantial
contribution for power supply of the future.
For Germany it is not possible to use tidal energy because of
technical reasons. At German coastlines the tidal ranges
amount to just under 3 meters.
The basic problems in using wave power are trivial and
generally known: Wave power is unstably and could not affect
in its size. Large waves have also a destructive nature.
The problem solutions, in order to arrange an economical
usage of wave energy with justifiable impact on the environment, are however not trivial.
It is very complex and expensive to build a power plant which
resists the large waves.
Researchers believe that only large power plants with a high
efficiency are competitive.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
335

As well as the potential of wave power is relative low in Germany because the waves and streams at the coastlines and
estuaries are too small for an economical usage.
Finally, against the background of more scarcely becoming
resources of fossil fuels much efforts will be made to use the
actual potentials in order that energy from the oceans make a
contribution, whatever to which amount, to the worlds energy
mix. [14], [15]

[6]
[7]
[8]
[9]
[10]

REFERENCES
[1]
[2]
[3]
[4]
[5]

Cruz, J. (2008): Ocean Wave Energy Current Status and Future

Perspectives, Berlin, Heidelberg: Springer-Verlag 2008, pp. 1-2, 93-94
[Link]/html/[Link], date: 25 February 2009
[Link]/exarb/imig/sonnewindwasser/
[Link]#343, date: 25 February 2009
Kaltschmitt, M., Streicher, W., Wiese, A. (Hrsg.) (2006): Erneuerbare
Energien Systemtechnik, Wirtschaftlichkeit, Umweltaspekt, 4. Auflage,
Berlin, Heidelberg: Springer-Verlag 2006, pp. 598-600
Giesecke, J., Mosonyi, E. (2005): Wasserkraftanlagen - Planung, Bau und
Betrieb, 4. Auflage Berlin, Heidelberg: Springer-Verlag 2005,
pp. 107-115

[11]
[12]
[13]
[14]
[15]

[Link]/EandE/Web_sites/01-02/RE_info/
Tidal%[Link], date: 26 February 2009
[Link]/publications/
survey_of_energy_resources_2007/tidal_energy/[Link]
date: 25 February 2009
Quaschning, V. (2007): Regenerative Energiesysteme Technologie,
Berechnung, Simulation, 5. Auflage, Mnchen: Hanser-Verlag 2007,
pp. 285-286
[Link]/23/advantages_of_marine_currents,
date: 28 February 2009
Giesecke, J., Mosonyi, E. (2005): Wasserkraftanlagen - Planung, Bau und
Betrieb, 4. Auflage Berlin, Heidelberg: Springer-Verlag 2005, p. 116
Giesecke, J., Mosonyi, E. (2005): Wasserkraftanlagen - Planung, Bau und
Betrieb, 4. Auflage Berlin, Heidelberg: Springer-Verlag 2005,
pp. 117-121
Quaschning, V. (2007): Regenerative Energiesysteme Technologie,
Berechnung, Simulation, 5. Auflage, Mnchen: Hanser-Verlag 2007,
pp. 286-287
Cruz, J. (2008): Ocean Wave Energy Current Status and Future
Perspectives, Berlin, Heidelberg: Springer-Verlag 2008, pp. 287-290
[Link]/exarb/imig/sonnewindwasser/
[Link]#343, date: 29 February 2009
[Link]/de/wasser/detailansicht/article/160/
[Link], date: 29 February 2009

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
336

Carbon Capture and Storage Part of the solution

to the climate change problem?
Robin Kuscheck, Sren Schlter, Martin Noack
Brandenburgische Technische Universitt Cottbus

I. INTRODUCTION
Focusing the state-of-the-art of renewable energy science and
technology, you are forced to advance the established energy
sources. Today, its not possible to substitute coal-burning
power plants, nuclear power plants or gas-fired power plants
for solar power stations or wind power stations.
The task is also to explore environmentally compatible
technologies for reduce the carbon dioxide emission and
uncouple it of the economic growth. The energy demand grows
because of the emerging markets like China or India and their
coal deposits. The Kyoto Protocol or the objectives of the
European Union are confronted the countries with the problem
of global warming resulting from the greenhouse gases.
II. CARBON CAPTURE AND STORAGE
For example, the Carbon Capture and Storage (CCS)
technology allows an environmentally compatible consumption
of fossil fuels. Before, during or after the power plants process
the carbon dioxide should be absorbed followed by
underground storage at the point sources. According to experts
and scientists the CCS technology will be fully developed
around 2020 and will be a part of the future energy mix.
The CCS is a process consisting of three elements: capture,
transportation and storage of carbon dioxide.
There are several ways for capturing: the pre-combustion, the
oxyfuel, the post-combustion.
During the pre-combustion a fuel (e.g. coal) will be extricated
before the combustion at power plant. At a high temperature,
the fuel is separated into synthesis gas of carbon monoxide and
hydrogen. In the next step, the carbon monoxide converts with
a help of steam into carbon dioxide, which can be sectioned out
of this gas mixture and be compacted into liquidity. This
procedure is also known as Integrated Gasification Combined
Cycle.
Air
Separation of air
Coal / O2
Gasification
Gas purification
Deposition of CO2
Power generation
At the oxyfuel, the proportion of oxygen in air is boosted to
reduce the fumes. Instead of nitrogen and sulfur compounds,
only hydrogen and carbon dioxide are the residues of the
combustion in oxygen followed by the cooling down and

condensation of water. The advantage of the oxyfuel is the

clean carbon dioxide after the burning in clear oxygen (see
below Schwarze Pumpe).
Air
Separation of air
Coal / O2
Boiler
Fume Cleaning
Deposition of CO2
Another possibility is the post-combustion with the expansion
of the conventional flue gas cleaning. After the denitrification
and desulfurization, the flue gas flows through a fluid of amine
which reacts with the carbon dioxide. After heating, the
carbon dioxide is available in a clear state that can be
compacted.
Air / Coal
Conventional Steam Power Plant
Fume Cleaning
Deposition of CO2
CO2
In the majority of cases, the carbon dioxide cannot be storaged
close to the power plant. It is necessary to transport the liquid
carbon dioxide in pipelines or by motor trucks or ships. The
storage locations which can be in run are utilized natural gas or
oil fields (Enhanced Oil Recovery), deep coal steams
(Enhanced Coal Bed Methane) or saline aquifers.
III. SCHWARZE PUMPE/GERMANY
In 2005 Vattenfall decided on building a 30 MW oxyfuel pilot
plant, with an investment of 50 Mil. , which is the first visual
sign of Vattenfall's project on CCS. This installation is located
near the existing lignite fired 1600 MW power plant in
Schwarze Pumpe, Germany.
On May 29th 2006 was the groundbreaking ceremony of the
Vattenfall pilot plant, which has been in operation from the
middle of 2008. The initial testing programme will run for
three years. Thereafter, the pilot plant will be available for
other tests. The plant is planned to be in operation for at least
10 years. The pilot plant is an important milestone to reach the
goal of commercial concepts for carbon capture and storage at
coal fired power plants by 2015 until 2020. It will be the first
pilot plant in the world to use the oxyfuel capture method. The
construction of the 30 MW thermal pilot plant at Schwarze

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
337

Pumpe is an important thing for the Vattenfall project. It is the

necessary link between initial engineering and successful
operation of the future 250350 MW electricity
demonstration plant. Vattenfalls goal is to develop a concept
at a total cost of approximately 20/tone of avoided CO2. But it
is to mention, that not only the plant in Schwarze Pumpe is
inaugurated to CSS.
Figure 1. Carbon Capture Unit

Further, nearby the plant in Jnschwalde, which uses no CO2

reducing system, Vattenfall currently investigates the
possibilities of implementing both oxyfuel and postcombustion technology at these plants.
The complete implementing could be realized in the year 2015.
IV. WORLDWIDE DEVELOPMENTS AND PROJECTS
The mentioned desire to use fossil fuels without polluting the
atmosphere implicates other projects in research and
development on CCS technology. The center of that can be
sight in Europe and the United States. Furthermore, developing
and emerging markets even show the readiness for cooperation.
For example, the public Algerian energy association and a joint
venture of international associations prove the separation of
CO2 in the Sahara Desert. In addition, India and China
cooperate with the Great Britain or rather with the United
States. First of all, these cooperations conduce to find storage
locations. The important role of China in CCS technology
because of the determining importance of climate protection
depends of the rapid economic growth, the growth of the
energy demand or the coal deposits.
Beside German companies, Great Britain and the United States
regard the CCS technology as a central topic of their energy
policy. For example, the US-Clean Coal Power Initiative
receives a two billion Dollar budget annually till 2011.
In Germany, the public research spending for the projects
Geotechnologien or Cooretec aggregate around 1.5 Billion
Euro for the whole energy und climate section from 2008 till
2011.

sub-process or whole the process chain of deposition of CO 2

and in questions of transport and storage.
The Australian Otway Basin Pilot Project wants to press big
quantities in empty natural gas fields as soon as to study the
other CCS technologies.
The US-CO2 Capture Project focuses the competitive
deposition of CO2 and the safety aspect of the three methods.
The European CO2 ReMoVe (research, monitoring,
verification) explores the criteria of the storage and
technologies of feeding storage locations.
The international Carbon Sequestration Leadership Forum
consists of 21 countries and supports the exchange of
information and multilateral collaboration.
V. CONCLUSION
The Carbon Capture and Storage technology seems to be a
trendsetting way in future energy mix, next to the renewable
energy. It is an advancement of conventional power plants,
which ensure the power supply. It is necessary to broaden the
existing international cooperation to reach the goal as soon as
possible.
But we are also forced to ask, if the carbon capture in natural
gas or oil fields, deep coal steams or saline aquifers is the right
solution. At the moment, we dont see possible risks in the faraway future. In order to guarantee the success of the project, it
is necessary that all countries which use coal plants introduce
the CSS technology. If this takes part of the energy production,
the benefits for the environment will be increased.
REFERENCES
Vattenfall AB:
[Link]
[Link]
CCS: IZ Klima Informationszentrum klimafreundliches
Kohlekraftwerk
e.V.
[Link]

The priorities and research objectives of the CCS projects

reached a big variety. Engineers and academics are engaged in

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
338

FFT algorithm optimization using Stream SIMD

Extension instruction set
Pawel Dawidowski
Wroclaw Uniwersity of technology
This paper explain how to perform Cooley and Tukey fast
Fourier transform for sequence length being power of 2 using
modern Stream SIMD1 Extension known as SSE. Basis of FFT
algorithm will be shown pointing on which parts can be
significantly speed up using SSE focusing on programming side
rather than FFT algorithm itself. Also basic SSE instruction will
be introduced.

I.

FFT

The idea of calculating digital convolution is to approximate

integral in (1) to (2) on finite sum over N samples of signal:

f (t ) e jt dt

(1)

N 1

X k = xn e

2 j
kn
N

k = 0,1,..., N 1

(2)

n =0

Point is this raw calculation need to perform N

multiplication over k which takes also N values so overall
complexity of algorithm is N2. Even for relatively small
numbers of samples the amount of calculation can be pretty
high. Assuming N is power of 2 Cooley and Tukey has split
odd and even elements of sum (2):

W kn = e
Xk =

N 1

2 j
kn
N

W kn +

n even

(3)

N 1

W kn

(4)

n odd

Each of those summations is recognized as being an N/2

point DFT of the respective sequence because:
2 j
2
N

SSE is a mathematical coprocessor which operates on

floating point numbers. It has 8 additional and independed 128bit registers for calculation which are 4x32-bit single precision
floating point value or 2x64 double precision one. The main
advantage of using it is performing 4 or 2 operation in parallel,
so depending on precision we need it can cut the algorithm
execution time 4 or 2 times.
Basic SSE instruction set is shown in table 1, while full
instruction set with descriptions and timing is available at Intel
web site (Ref. [2]). All instruction which end PS is doing 4
parallel operations at simultaneously while SS mean only one
operation.
TABLE 1
SSE BASIC INSTRUCTION SET
Instruction
MOVSS,
MOVAPS, MOVUPS
MOVSD,
MOVAPD,MOVUPD
ADDPS, ADDSS,
ADDPD, ADDSD
SUBPS, SUBSS,
SUBPD, SUBSD
MULPS, MULSS,
MULPD, MULSD
DIVPS, DIVSS,
DIVPD, DIVSD
SHUFPS

Operation
Copy operation:
from memory. to register
from register to memory
from register to register
Addition
Subtraction
Multiplication
Division
In register data shuffle

2 j
N
2

W2 = e
=e
(5)
So if we calculate DFT for odd and even indexed k in (2) we
can calculate overall DFT combining (4) together:
(6)
X (k ) = X even (k ) + W k X odd ( k )
Using this approach we will need only N log2N complex
multiplication to perform, because we need exactly log2N of 2
times smaller DFT to calculate.
For example if N = 1024:
DFT : multiplication need = N 2 = 1048576
(7)
FFT : multiplication need = Nlog 2 N = 10240
Because of this great time reduction FFT algorithm is widely
used in many application which need Fourier transform.

II. STREAM SIMD EXTENSION

SIMD is abbreviation from Single Instruction Multiple Data. In the term of

SSE instruction 4 single precision floating point or 2 double precision floating
point operation are performed in parallel on one processor core.

There are few things which can be optimized using those

instructions. First thing is to rewrite complex multiplication to
take advantage of parallel operation. Second is to notice, that in
first step Wkn = -1 so no complex operation is actually
necessary. In second step we have at most 4 Wkn values 1, j, 1 ,-j which can threaten as a special case. Since Wkn are kth root
of one we can calculate it inside of SSE register for next steps.
Only value need to load is to load complex number for k = 1 in
(3). Next values for next k can be calculated by recursion (8):
(8)
W kn +1 = W kn W 1
III. COMPLEX MULTIPLICATION ON SSE
To calculate complex multiplication using SSE we can
assume at the beginning that we want to perform 2
multiplications at once. With 4 floating point register each pair
of 2 values can be complex real and imaginary part. For start
assume we want to calculate:

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
339

IV. MEMORY OPERATION FROM AND TO SSE REGISTERS

Z A = a A + jbA
Z B = aB + jbB

(9)

ZC = Z A Z B
From algebra we know:
Z C = a A aB bA bB + j ( a A bB + bA aB )

(10)

We need 4 multiplications and 2 algebraic additions to

calculate complex multiplication.
Now lets try to do this using SSE. Since we have 4 floating
value in one register we can perform 2 multiplications in
parallel. Lets assume we have loaded our complex numbers
into XMM0 and XMM1 registers and XMM3 and XMM4 has
values to change signs. Each of them contains 2 complex
numbers as in the Table 2:
TABLE 2
INITIAL VALUES IN REGISTER
Register name
XMM0
XMM1
XMM3
XMM4

reg[0]
aA1
aB1
1.0
-1.0

reg[1]
bA1
bB1
-1.0
1.0

reg[2]
aA2
aB2
1.0
-1.0

reg[3]
bA2
bB2
-1.0
1.0

The program which realizes complex multiplication is:

1.
2.
3.
4.
5.
6.
7.
8.

MOVAPS
SHUFPS
MULPS
SHUFPS
MULPS
MULPS
SHUFPS
ADDPS

%%XMM0,%%XMM2
$0xA0, %%XMM2,%%XMM2
%%XMM1,%%XMM2
$0xF5, %%XMM0,%%XMM0
%%XMM3,%%XMM0
%%XMM1,%%XMM0
$0xB1 %%XMM0,%%XMM0
%%XMM2,%%XMM0

Table 3. shows how register has changed as a result of

operation:
TABLE 3
REGISTER VALUES AFTER OPERATION
Line
1.
2.
3.
4.
5.
6.
7.

register
XMM2
XMM2
XMM2
XMM0
XMM0
XMM0
XMM0

8.

XMM0

reg[0]
aA1
aA1
aA1 aB1
bA1
bA1
bA1 aB1
-bA1 bB1
aA1 aB1 bA1 bB1

reg[1]
bA1
aA1
aA1 bB1
bA1
-bA1
-bA1 bB1
bA1 aB1
aA1 bB1 +
bA1 aB1

reg[2]
aA2
aA2
aA2 aB2
bA2
bA2
bA2 aB2
-bA2 bB2
aA2 aB2 bA2 bB2

reg[3]
bA2
aA2
aA2 bB2
bA2
-bA2
-bA2 bB2
bA2 aB2
aA2 bB2 +
bA2 aB2

As we can see we need only 3 parallel multiplications and one

parallel addition to calculate 2 complex multiplications at once.
It means we need only 1.5 multiplications not 4 and only 0.5
additions not 2 per one complex multiplication. So
optimization has reduced floating point operation from 6 to 2.
Thats 3 times faster than without use SSE. All copy and
shuffle operation are only on internal registers which is done
much faster than copying values from/to memory.

There are two main instructions to load from memory into

XMM register and to store XMM register into memory:
MOVUPS and MOVAPS. First one work as a normal MOV
operation but it will copy 4x32bit floating point number into
XMM register. MOVAPS is faster but with one constraint: all
memory addresses has to be 16 bytes aligned, it means four
least significant bits in address has to be equal zero. The speed
difference in practical application is about 20-30% with
MOVAPS in comparison to MOVUPS. Luckily for FFT
calculation every step will be 16 bytes aligned since overall
data length is power of 2 so only thing to do is enable 16 bytes
align in compilation options.
V. OPTIMIZING 1ST AND 2ND FFT STEP
For 1st step W = -1 so there is no need to perform any
complex calculation. For second step W = 1, j ,-1, -j. For those
special cases we can write even shorter program. Only 3th and
next steps will need full complex multiplication.
Multiplication by j is only shuffling real and imaginary part
and change the sign properly as shown in (11) so it can be
realized on 2 SSE instructions for 2 complex numbers in
parallel.
(11)
Z j = ( a + jb) j = b + ja
The program which realize complex multiplication by j
using initial values as in Table 2 except XMM1 and XMM4 are
skipped:
1.
2.

SHUFPS
MULPS

$0xB1, %%XMM0,%%XMM0
%%XMM4,%%XMM0

The point is that even if we want to swap real and imaginary

part it could result in multiple memory read/write operation.
Here 2 complex numbers are read from and write to memory
exactly once so overall floating point operation number is 0.5
multiplications per one complex number.
VI. SUMMARY
Well optimized FFT algorithm can calculate Fourier
transform online, even for quite large samples number, while it
would be impossible to use raw DFT algorithm to calculate it
even on modern processors. Using SSE cuts computation time
even more by taking advantage of parallel calculation and
coprocessor designed especially for floating point operation.
All instructions presented in this paper are available in C++
either by inline assembly code or by intrinsic functions fully
described on Microsoft web page (Ref. [3]).
VII. REFERENCES
[1] Leland B. Jackson Digital Filters and Signal Processing , 3th ed.
[2]
[3]

University of Rhode Island. Kluwer Academic Publishers Boston

Dordrecht London
[Link]
[Link]

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
340

!"!#$%&'&( )
*
+,-, . /0 !'!%!1!%$
) 0 23$' 4!5%%4##%(6
/0 23$' 4!5%%4##%(! )

,*
,

:
:

@
:
;

="> =!6>, *
;
: ;
:
=!4>,
;

=!!>,
/

'

;
;
;

&

=&>,

;
:

"# $%

:
;

;
:

;
;

;
)

;
;

;;
*

,
: ;

=!&>,
; ;

;
,.

;
*,

: ;

*7) -8 9)*-7
:

/ ;

-+

$
/;

, 9
;

;
<

, )

, A

,)

;
B;

=!>,
/;

:
: ;
;

/;
;
/

? :

:
2!$$'

9
;

**,

=4>,

; ;

!$$$5,
:

) +-A

;
;

:
,)

:
:
=!">,

: ,
:

;;

;
: ,-

:
:

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
341

;;

4
;

/;

, A

2!5

: ;
;
;

, +

;
;

<

245
,

:
;
;

: ;

=!%>,

; :
;

;;
;

, *
;

; :
;;

,
;
;

;;
;

;
:
;

, )
;

;
; :

;;

;
;

!
, .

;
;

;
!

;
,.

2&5
=!(>, 9
:
:
: ;
,
:
@

/;
*

;;

, 8

;;

;
2/5

C
D
;
=!#>, A

0
2"5

:
;
) E + ;
A
E +
;
:
) E +

<

@
;

/;

;
:,
;;

, *

*J,

:
:

0
5 3 2

E + F 2
2

#,!
)+-A
,)
(

5 3

5,
)

:
, )

E )*-7

,
-+

&,4
$

,)

!0

;
E +

***,
)

9
)+-A
,)

9 -+

;
G

E )*-7

;
H -+

<
=!'>, )
;

;
; ,)

!, $

1
)

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
342

&
)
K
!
4
&
"
%
(

!, )
.
!
4
&
%
'
$

;
2 A5

+
!6
!6
!6
!6
!6
!6

;
+ /2 A5
4%6
&66
4#6
4%6
&66
&66

;
;

/,
;

4!,&'#

4!,"#'

44,!#&

4!,&'#

44,64&

- ; 1 A

#6,66

!4(,&%

#%,66

#6,66

!4(,&%

!%%,66

(44,#6

!6&$,66

4!6',%"

!6$$,6(

!6&$,66

4!6',%(

4'4',!4

!6444,4$

1 2L1 5

%4%,4(

%$&,#4

%!!,#$

%!&,!6

%$&,#4

%'%,%!

&&4&,!!

12L1 A 5

4#,666

4!,&'#

4&,44%

4%,'!6

4!,&'#

46,#""

- ; 1 A

!6%,4&

!"%,%'

#%,66

#6,66

!"%,%'

'4,%6

(4&,$6

!'$",&"

4%#(,!'

!6$$,6(

!6&$,66

4%#(,4"

!4##,%&

!6"(4,&%

$"(,$$

%&#,""

("4,'"

#(#,#4

%&#,"&

"&&,'(

&'((,4'

"4!,#&

%(,4'

!&!,6%

4%",(4

%(,4$

;
, *

1 2L1 5
1 2L1 5

:
;

! &

* ;
E +

,)
;
4,

"

C
;
;

E +

B;
)

!%!,(%

%"&,!#66

;
;
;
!6%,4& A,
,
:

#6,66 A
;

;
;

K!

4"6,6(L1
%"&,!#L1 , *
, K

C
;
,)
E +

;
K(

40

44,&"#

1 2L1 5

E +

$
K%

1 2L1 5

"

B;

K"

K&

" $ #$,' A, )
"6 A

K4

E
"$

K!
12L1 A 5

;
;

4, -+
9

;;

,!

@ ;

: ,
;

,A

0
!
$ %& '

,*
E +

"
(

#
#
;

;
, *
;

!
;
: ;

, K!

;
;

, )
&0

4, ;

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
343

"
)
12L1 A 5
- ; 1 A
5 L1 1 2 9
5 L1 1 2 +
E$ "

&, K!

6
44,&"#

!6
4#,!(4

!6,%
4',"#6

!!
4',%6!

!!,%
4',%$&

!4
4',#6&

#6,66
!6&$,66
%4%,4(

"(,'%
#6!,4%"
%#!,4'
9

"',4%
%'%,'!4
#'#,'(
9

&%,4!
"#&,4(!
%6&,4(
9

4#,'(
&"(,''%
""$,#!
9

!#,''"
!'',$%&
&4",&#
9

)
K!

&

", K! ;

;
E +
2. !5
, A
E + 44,&"#
L1 A
K!@
;
#6 AD
: !4
E +
4',#6& L1 A
K!@
;
!#,''" A, .
;
K! ;
E +
;
;
K!@ ;
%4%,4(L1
#'#,'(L1 2 /
:
!6,%5
,A
:
!4
E +
K!@
;
K!@ ;
%4%,4(L1
;
K!
;
,A
: !6 E " $
,

&, K

B;

$ )& *

"
;
;

;
K!

,A
;
;
E +
, K!@ ;
%4",6" L1
#"","%L1 2 /
K!@
;
%6 A5
,A
;
2 ; 5
(6 A E " $
,
)
&
)
"
;
;
, A
:
!6
K!@
;
2 ; 5
"(,'% A E
"$
,
A
;
2
: 5
;
2
K!@
;
5
;
,

"

K!
;

@
;

E +
)
",
;
/
/

E
;
;
;
;
M

;
A

K!@
"$

2. M !5
;

#6

;
;

%, E

/
;

K!

K!

4%6 A, *
;
M
#6

A
/
A

K!
K!@
;

46

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
344

%
)
.

1 A

", K!@ 8 ) A*)I A*)II-E8*7K 9 + 9*)N

(6
%%
%6
"%
"6
&6

#6

46

- ; 1 A

#6

(6

%%

%6

"%

"6

&6

46

12L1 A 5

44,&4$

4%,#64

4%,#6$

4',&'$

4',"&(

4',"'"

4',%#!

4',(#&

!6&$,6!

'"(,6!

#%#,#(

(#%,6!

%$#,#(

%4(,6!

&$$,6!

4$",66

%4",6"

((6,!&

(%(,4&

#"","%

('!,'$

(!&,&#

"%',!&

4#$,"#

12 L1 5

5L1 1 2 +
E$ "

;
2
;

79

,A
K!
/
;
#'#,'(L1 2
:
"',4% A5, A
K!
/
;
#"","%L1 2
%6 A5,
,
;

!6,%
K!@

*-7

;
:

:
:

E +
, *
/
;

,
J*,

%,K

++ 78*O

B;

O6

N6

O!

N!

O4

N4

K!

!%66

&

6,!!

!%6

K4

4666

&

6,6'%

!,4

(66

K&

&666

&

6,!44%

&&%

K"

!%66

&

6,!!

!%6

K%

4666

&

6,6'%

!,4

(66

K(

4666

&

6,6'%

!,4

(66

;
0
/6 6 /! ! /4 4 ,,,
A
/6 P /! P /4 P ,,,
2/4 45 ,,,

;
:;

7 ;
* ;
H *
.

+
J-E, 44 7-, !
=4> ?
.
Q)
9
*
; ;
+
=&> K
2466(5 Q*

, 7

.
8

="> N

J, 9-79E
)

=!>

;
K!
K!@
;
;
;

2/6 65 2/! !5
,

K, ?

) 7
N 466#
;

K
.
.
9)*-7 -7 +-A

!$$$ )
A :

G)
N )

: +
Q
; ;
;

:
2+-A

K
Q

Q =

5
>,

/
: ;
Q
),,,(
#(
+ ;;,44!# 4444,
=%>
E,
Q
: ;
0
Q
% .
/
(#
0#
1
23
1
456 8,K
8, J,.
Q E
: ;
0 /;
Q'
7
(#
, !# ;;, !'6 !'%
, 4664,
=#> K
.,
Q 9 ;
;
;
Q
+
!$$$,
='>
< *
K
E
:
;
0
Q
+
!$$$,
=$> 8, K
8, J,.
Q
;
: ;
Q
),,) (
# *
;;, "&"M"&$,
=!6> ,
E,8,., )
Q
: ;
Q
%
1
;;,'4,
=!!> ,
9
,A
?, K
'
#, 7
N :0 -/
, +
!$$%,
=!4> ,
(#
0
/
,7
N :0 *
1A
4664,
=!&> 9
+,
+
K
?,
+ 8.
8
8
) ;
+
: +
;
H *
) 7 9)*-7 -7 +-A
N )
J-E, !$ 7-, 4
N
466"
=!"> 8
,
, Q
:
+
;; Q *
)
J ,!( 7 , " +
0 &%4 &%&
8
466!,
=!%>
., 9
E,
, Q
;;
+
: +
9
: Q *
+
&(
7
+
;
$ !6 466"
*
*
,
=!(>
? )
.
9, E
)
8,
,
Q*
; :
: ;
;
8
;;
Q J , "6 *
&" 466%
+
0 %!# %4' * 706!(# $4&(,
=!#> 8
K
G
+
H ,--. ?
A
E ,
=!'>
8, R
8 C
K ,
;
@
4"
STTU,

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
345

Banning of Incandescent Bulbs and Its

Influence to Electromagnetic Compatibility
LU Yan

AbstractThis paper presents worldwide Banning of

incandescent bulbs campaign, which aims to replace traditional
incandescent bulbs that have long been known as inefficient
sources of electrical lighting, with compact fluorescent lamps
(CFLs). Controversy, operations and comparisons between these
two alternatives are introduced in succession. Electromagnetic
compatibility (EMC) issues, such as light flicker, which emerge
more apparently because of the coming popularization of CFLs,
will be discussed. Additionally, precautionary senses and steps to
limit adverse interferences are mentioned when installing
lighting systems.
Index Termsincandescent light bulbs, compact fluorescent
lamps, flicker, electromagnetic compatibility

I. BACKGROUND

he worldwide Banning of incandescent bulbs

campaign has been supported by most governments,
according to pass measures to phase out incandescent light
bulbs. In some jurisdictions this has been done through
legislation (Cuba, Brazil and Venezuela phased out
incandescent light bulbs in 2005); while others, like United
Kingdom, through voluntary measures. The aim is to
encourage use of more energy efficient lighting alternatives,
such as compact fluorescent lamp (CFLs) and LED lamps.
For European Union, the Irish government was the first
member state to ban the sale of incandescent light bulbs. It was
later announced that the member states of the EU agreed to a
phasing out of incandescent light bulbs by 2012, which will
save the European economy up to 10 billion euros.
Nevertheless, it is up to the government of each member state
on how to accomplish the eventual phase out. For example Italy
will accomplish this through a ban on their sale, while the UK
has enlisted the help of retailers with a voluntary, staged phase
out.[1][2][6]

glass tubing, averaging 4.0 mg per bulb. A broken compact

fluorescent lamp will release its mercury content. Safe cleanup
of broken compact fluorescent lamps differs from cleanup of
conventional broken glass or incandescent bulbs. Because
household users in most regions have the option of disposing of
these products in the same way they dispose of other solid waste,
most CFLs are going to municipal solid waste instead of being
properly recycled.
B. Cost and existing fixtures
The cost of CFLs is higher than incandescent light bulbs.
Typically this extra cost may be repaid in the long-term as
CFLs use less energy and have longer operating lives than
incandescent bulbs. However, there are some areas where the
extra cost of a CFL may never be repaid, typically where bulbs
are used relatively infrequently such as in little-used closets and
attics. Use in situations such as stairways where the lamp is
turned on for less than five minutes at a time will cause a
significantly shorter lifespan for a CFL. In the case of a
5-minute on/off cycle the lifespan of a CFL can be up to 85%
shorter, reducing its lifespan to the level of an incandescent
lamp.
Typical CFLs may have issues when being used in older
dimmer light fixtures and with some electric timers, meaning
many people would need to replace fixtures to retain that
functionality. However some specialized CFLs can be
purchased which may fit into the fixture without the need for
replacement.[6]

II. CONTROVERSY OF THE CAMPAIGN

A. Environmental concerns
As the recommended alternative, CFLs, like all fluorescent
lamps, contain small amounts of mercury as vapour inside the
_______________________________
LU Yan is with the Faculty of Electrical Engineering, Wroclaw University of
Technology, Wroclaw 50-370, Poland (e-mail: yanr61@[Link]).

Figure 1. Interior configuration of an incandescent bulb

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
346

III. OPERATIONS AND COMPARISONS OF INCANDESCENT

BULBS AND CFLS
Incandescent lamps, with the interior configuration shown
in Fig. 1, consist of a tightly coiled tungsten filament supported
in an inert gas typically argon encapsulated in a glass bulb.[6]
Electrical current passes through the filament heating it to
2000-3000K where it emits photons in the visible range as
well as considerable amount in the Infra-Red range.
Fluorescent lighting excites gaseous mercury atoms in a
plasma arc. These atoms release photons in the UV range
which then excite a phosphor coating to produce visible light.
The plasma arc exhibits negative resistance and therefore
requires an electrical ballast to limit the current.
Traditional incandescent lamps have remained relatively
unchanged since the invention in the 19th century and have
long been known as inefficient sources of electrical lighting.
With current global pressures towards energy efficiency, CFLs
offer 4-5 times the efficiency using fluorescent technology.
From the statistical result of Fig. 2, we can get the
straightforward understanding of efficiency discrepancies.[6]
By compacting the discharge tube and developing low cost
electronic ballasts, domestic use is increasing.
However, a number of concerns have been raised when
replacing incandescent lamps with CFLs. Some lamps: 1)
produce a 'cool white' colour not matching the 'warm white'
colour of incandescent bulbs, 2) have delayed start-up and/or
take considerable time reach full brightness, 3) fail to produce
the equivalent rated light output, 4) produce audible sound, and
5) are usually non-dimmable. Aside from these consumer
aspects, CFLs like incandescent lamps are susceptible to power
system disturbances resulting in light flicker.[3]

Figure 2. Energy usage for different types of light bulbs

Figure 3. Typical CFL ballast design

IV. INFLUENCE TO THE ELECTROMAGNETIC COMPATIBILITY

Light flicker (Flicker) is the variation of luminous flux (or
visible light) within the frequencies perceptible by people. It
generally has a adverse physiological effect on persons causing
eye strain and/or even annoying emotion. In rare cases it can
have serious health risks such as being stranded in distraught
situation, loss of concentration, headaches or the triggering of
epileptic seizures.[3][6]
However, because CFLs use electronic ballasts, their
subsequent non-linear behaviour has made modelling of their
operation difficult. Equally their susceptivity to voltage
disturbances resulting in visible light flicker is widely
unknown and difficult to predict. This has given rise to
discrepancies between the measured flicker levels by relevant
international flicker standards and the actual flicker produced.
A typical CFL ballast[5], as shown in Fig. 3, consists of a
single phase rectifier stage creating a DC bus from which a
high frequency inverter resonates to control the fluorescent
tube current. Due to the diode rectifier, some require EMC
filtering to comply with national harmonic standards, some
however omit this. During starting two filaments at either ends
of the tube are heated to create enough free ions for the plasma
to be formed. Depending on the designed heating time, a
delayed start-up can occur, on the other hand, insufficient
heating reduces the lamps life.
For an uncontrolled full bridge rectifier the AC side voltage
relates to the DC bus voltage through the switching function
(modulation) by the diodes. Fig. 4, shows a generalised case of
a rectifier switching function on a voltage waveform with a
small distortion. Given that an AC side disturbance results in a
modulated DC disturbance, the light output is likely to
experience a similar fluctuation and result in visible flicker.[3]

Figure 4. Switching operation of uncontrolled single phase rectifier

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
347

V. INSTALLING PRECAUTIONS RELATED TO EMC ISSUES

The lighting system is one of the most pervasive of the
electrical loads. Lighting can create localized electrical field
problems and affect susceptible electronic equipment.[4]
Especially for the coming popularization of CFLs, problems
related to electromagnetic compatibility emerge apparently.
Electric fields develop around fluorescent lighting fixtures, and
today's electronic ballasts create high-frequency fields. The
impact of these fields depends upon the sensitivity of
equipment within those fields and the distance from the
fixtures to the equipment.
Obviously, installation engineers should consider adverse
interference effects when making selection of lighting fixtures.
Our advice is to control the end results in the beginning rather
than spend countless dollars and hours combating problems
after the fact.
To minimize interference effects, take the following steps:
1. Install all wiring for lighting in separate conduits and
power lighting circuits from dedicated-use isolation
transformers.
2. Never combine lighting circuit wiring with computer
system or other susceptible electronic system wiring in
the same raceway.
3. In areas with susceptible equipment, use lighting
systems without high-frequency electric fields (e.g.,
DC) or lighting fixtures with electromagnetic
interference screens to control the electric fields.
Also, there are a few simple precautions should be taken for
installing lighting systems near computers and data cables.
They aren't new, and can be found in numerous site preparation
manuals. Nevertheless, they warrant repeating here:[4][6]
1. Keep lighting circuits out of computer power panels.
2. Do not install data cables so that they run on top of
fluorescent lighting fixtures because the magnetic
fields from electronic ballasts may inductively couple

3.

into the data cable.

Make sure that you do not install poorly shielded data
cables in the electric fields below lighting fixtures.
VI. CONCLUSION

Banning of incandescent bulbs is the governmental

campaign to phase out the inefficient incandescent bulbs with
compact fluorescent lamps (or LED lamps), which are not only
taken into account of improving efficiency, but also should be
considered of its following influence to electromagnetic
compatibility issue, such as light flicker.
Because of CFL ballast, the induced flicker is much more
complex, which may cause serious healthy problems to
customers and subsequential detriment to the whole systems
power quality. To minimize those potential interferences,
installing precautions should be noticed and checked as the
indispensable condition.
VII. REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]

Georges Zissis: CFL Quality and Strategies to Phase-out Incandescent

Lamps. International Energy Agency, Paris (France), February 16th, 2007
Beletich Associates: Phase-out of Inefficient Incandescent Lamps and
Standards for Compact Fluorescent Lamps. Prepared for Australian
Greenhouse Office, December 2007
L. P. Frater, N. R. Watson: Light Flicker Sensitivity of High Efficiency
Compact Fluorescent Lamps. Proceeding of Universities Power Engineering
Conference, Australia, December 2007, page 1-6
Tom Shaughnessy: Lighting System and Electromagnetic Compatibility.
PowerCET Corporation, Calif., USA, May 2007
Eur Ing Ir David Cogan: Technical Specifications for Compact Fluorescent
Lamps with Integral Ballast. APEC Energy Efficiency Standards
Wikipedia website: [Link]

VIII. BIOGRAPHIES
LU Yan was born in 1984 in China. He received his [Link]. degree in University of
Electronics Science and Technology of China, Chengdu, in July 2007. From
October 2007, he started his master study of Control in Electric Power
Engineering in Wroclaw University of Technology, Poland. Currently, he is taking
one-year BOND project with AREVA T&D Ltd in UK. His interests are power
quality and artificial intelligence.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
348

Reactive power in one-phase circuits with periodical

voltages reactive power compensation
Sebastian Slabosz
Wroclaw University of Technology, Institute of Electrical Engineering Fundamentals
Janiszewskiego 8 Str., 50-370 Wroclaw, E-mail: [Link]@[Link]
Abstract - The article states about reactive power
compensation methods for circuits with non-sinusoidal voltages,
there have been presented selected theories application in order
to compensate the reactive power in one-phase circuits. Also
measurement results after compensation of an actual object
supplied from an non-sinusoidal voltage source were presented.
Also algorithms of optimal capacity selection were given, which
connected in parallel to the circuit with inductive character will
cause current root-mean-square value minimization.

I.

INTRODUCTION

Significant part of electrical energy receivers draw active

energy and convert it for work and heat, as well as reactive
energy. Reactive power describes the process of electrical
and magnetic field energy exchange. The process of energy
exchange can proceed independently from work and heat
exchange process. The utilization measure of energy supplied
to receiver is the power factor:

P
PF =
S

(1)

where: PF power factor, P active power, S complex

power.

Q = UI sin j

(3)
Basing on above mentioned dependence (3) it is possible
to select the optimal capacity with which we will achieve
power factor correction:

C=

Q T
U 2 2 p

(4)

Whereas for non-linear circuits or circuits supplied with

distorted voltages many power theories exist. In this article
selected reactive power theories will be demonstrated and used
to compensate an actual object (choke with ferromagnetic core)
supplied from public network.
III. REACTIVE POWER IN CIRCUITS WITH NON-SINUSOIDAL
VOLTAGES

Most disseminated reactive power theory for non-sinusoidal

voltages and currents is the one defined in 1927 by
C.I. Budeanu [1]:

Q = U n I n sin j n

(5)

n =1

As complex power we name the biggest active power value,

which can occur for given root-mean-square value of voltage
and current:

where: n harmonic number, j n - phase displacement

between voltage and current of the n-th harmonic.

S = UI

From the beginning the reactive power theory proposed

by Budeanu, besides many adherents, had many opponents.
The main charges for this theory are lack of physical
deformation power interpretation and unauthorized oscillatory
components summation of particular harmonics. In 1987
L.S. Czarnecki has published a scientific work in which
he criticized the Budeanu power theory [2]. He advanced
a hypothesis about its worthlessness for the sake of lack of
possibilities to:
- minimize the reactive power and therefore
no possibility to correct the power factor,
- reactive power QB is no measure of energy oscillation,
- it doesnt enable the calculation of capacitor capacity,
by which the power factor is the biggest,
- it
suggests
incorrect
energetic
phenomena
interpretation with distorted voltages.
One of the advantages arguing for this theory usage is its
conservatism and relatively simple measuring devices
construction. However this argument, in actual signal

where: URMS voltage root-mean-square value, IRMS

root-mean-square value.

(2)
current

If the receiver works with small power factor, it means that

it draws bigger current than it is necessary. If the reactive
power is drawn from distant sources it causes a rise of supply
currents and in consequence rises the transmission losses.
As a result it is needed to increase the transmission line
conductor cross section and lowering the ability to load the
generators and transformers with active power. All this aspects
are causing increase of operation costs incurred by the
electrical energy recipients.
II. REACTIVE POWER IN CIRCUITS WITH PERIODIC VOLTAGES
Reactive power is correctly defined for linear circuits with
sinusoidal voltages (3):

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
349

processing theory state, is insufficient according to the article

Author.
In the twenties of 20th century M. Iliovici has presented
a reactive power interpretation as loop area which is made by
current and voltage coordinates [5].

Q=-

1
idu
2p

i (t ) = ia (t ) + ib (t )

(6)

Characteristics in I, U coordinates of non-linear objects are

usually complex and create multiple loops and furthermore
their shape changes strongly under the influence of voltage
change. Areas inside loops are circulated clockwise
or counterclockwise. Therefore the energy of electric
or magnetic field is sometimes drawn and sometimes returned.
We select the capacitor capacitance connected to the object
to achieve a state in which the resultant loop area will be equal
zero, and so the reactive power will be equal zero. But
it doesnt mean that magnetic and electric field energy wont
be mentioned. If energy is not mentioned the object
characteristic in I, U coordinates is reduced to a line segment.
The reactive power compensation basing on above mentioned
theories is made using the formula (7), for given supplying
voltage we can determine an optimal capacity, which
connected to circuit (Figure 1) will cause rms value decrease.

C opt =

Fryze thought that power is a too elementary concept to define

it with assistance of such complicated instrument like Fourier
series, which was introduced by Budeanu. In his theory Fryze
distributed current to sum of two, mutually orthogonal,
currents:

2 p Q
u& 2 T

(7)

(9)

where: ia(t) active current component, ib(t) reactive current

component.
And so the reactive power was equal (10):

where:

QF = U I b

(10)

ia (t ) = Geu (t )

(11)

Where Ge can be determined from (12):

1
u (t )i(t )dt
T 0
T

Ge =

1 2
u (t )dt
T 0
T

(12)

An unquestionable advantage of Fryzes theory is elimination,

where: Q reactive power, u& - voltage derivative rms value, form initial Budeanu theory, of fourier series and third power
component (deformation power).

T time period.

i(t)
R
u(t)

For many years there hasnt existed a generally accepted

theory describing energetic properties of circuits with
non-sinusoidal voltages. The problem gathered meaning with
power electronics development. In seventies of twentieth
century Shepherd and Zakikhani have presented their reactive
power idea [7]. This theory is limited only to one-phase
circuits. The authors have distributed current sources into two
components:

i (t ) = iR (t ) + ir (t )

(13)

where:

iR (t ) = 2 I n cosj n cos(nwt + an )

(14)

n =1

Resistive current component,

Figure 1. Parallel compensation.

In 1931 S. Fryze proposed reactive power determination for

distorted voltages in such a way that the power equation
is fulfilled (8) [4], [5]:

S = P2 + Q2

(8)

ir (t ) = 2 I n sin j n sin( nwt + an )

n =1

Reactive current component, where:

an - argU n ,

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
350

(15)

j n - angle (Un,In)

Copt = -

Whereby these currents are mutually orthogonal:

(t )ir (t )dt = 0

(16)

The current defined by dependence (17) is called the passive

reactance current and can be interpreted as current which
is connected with reflexive source-receiver energy flow, and
his measure is the reactive power Q.

ir (t ) = 2 Re j 0 BnU n exp( jnwt )

(17)

n =1

This current can be compensated for finite harmonics number

with a reactive two-terminal network connected parallel to
receiver (Figure 1) with susceptance for every examined
harmonic.
This property was noticed for the first time by E. Emmanuel
[3]. On the basis of Shepherd and Zakikhanis theory it was
possible
to
determine
compensatory
capacitor
capacitance -optimal capacitance, for which the source
coefficient is the biggest:

Copt =

nU
n =1

I n sin j n

w n2 U n
n=1

IV. REACTIVE POWER COMPENSATION

In order to present the compensation effect with the help
of proposed by author theories, a choke with unknown
parameters was supplied with net voltage as shown on
(Figure 2):
400
300
200

0
-100

-300

nQ
n =1

100

-200

(20)

Compensation made with this method is a lot easier than

according to Shepherd and Zakikhanis theory. Also the
capacitance values calculated according to formula (20) are
identical like for dependence (18).

Voltage [V]

u& o i
Q
=
2
du
u&
u
dt

-400

w n2 U n

0,005

(18)

0,01

0,015

0,02

Time [s]

Figure 2. Voltage versus time graph

n =1

The advantages of this method are:

- ir(t) current determination, which can be compensated
for finite harmonics number using reactive twoterminal network,
- optimal capacitance value determination.

For voltage rms value Urms=235.65 V the system response was

current shown on (Figure 3) with current rms value
Irms=2.7511 A. The power factor value was PF=0.127.
20
15

i = ia + irC + irCs

(19)

Optimal compensation capacitance is then defined in the

time domain as:

10
Current [A]

Kusters and Moore have created a theory which enables

to calculate the optimal compensating capacity [6], which
significantly simplifies the compensation in comparison with
Shepherd and Zakikhanis theory, because of its harmonics
manipulation, which requires the knowledge on phase
displacement of particular harmonics.
According to this theory the current can be resolved into
three orthogonal components: active current ia (identical with
active current from Fryzes definition), reactive capacity
current irC and complementary reactive current irCs.

5
0
-5
-10
-15
-20
0

0,005

0,01

0,015

0,02

Time [s]

Figure 3. Current versus time graph

After making the observation of current it can be seen that

the element is a strongly nonlinear object. Reactive power
according to C.I. Budeanu, evaluated from (5) equation, equals
Q=550.59 var. Optimal capacitance, which has to be connected
to the circuit to achieve the compensation, was evaluated

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
351

according to (4) equation, equals Copt=2.40 mF. Within making

the measurements this capacity was corrected depending on the
value of reactive power; more and more the reactive power was
closer to zero, the capacitor capacitance was determined more
precisely. The current rms value was Irms=1.3680 A and the
power factor was PF=0.177.
Performing the reactive power compensation supported by
the M. Iliovici [5] theory, for the same measurement
conditions, the reactive power calculated from the loop area
which is made by drawing the u=f(i) function (Figure 4) equals
Q=546.48 var; and the optimal capacitance equals
Copt=2.34 mF. Within making the circuit compensation the
capacitance value was corrected. After the circuit
compensation the current rms value was Irms=1.3601 A and the
power factor was PF=0.189.
For comparison the reactive power value according to
C.I. Budeanu was Q=15.422 var. This means that making the
compensation according to (5) the circuit is overcompensated.
400
300

Voltage [V]

200
100
0
-100
-200

Performing the reactive power compensation supported by

the Kusters and Moore theory for the same measurement
conditions, the optimal capacity Copt=36.04 uF was determined
according to (20) dependence. After making the compensation
the current rms value was equal Irms=2.5925 A and the power
factor was equal PF=0.128. It means that the compensation
made using the dependence (20), proposed by Kusters and
Moore, the circuit will be undercompensated.
V. CONCLUSIONS
Taking advantage of mathematical dependences and modern
measurement techniques we can match the compensator
capacitances, without knowledge on receiver parameters. It is
a very important practical feature because the compensation
can be made in real-time. Thanks to this it is possible to make
quick changes in capacity value and its adjustment to receiver
work conditions and supplying voltage variations. The best
compensation of reactive power, this means the biggest power
factor is obtained taking advantage of M. Iliovici power theory.
While making the compensation using the C.I. Budeanu and
Kusters and Moore theory the circuit is overcompensated
or undercompensated and therefore current rms value increases
and the receiver operates with small power factor. Making use
of power theory utilizing the loop area calculation it is
possible, as shown it the article, to compensate strongly
nonlinear receivers supplied with distorted signals. It is a very
important practical feature.

-300
-400
-20

-15

-10

-5

10

15

REFERENCES

20

Current [A]

[1]

Figure 4. Reactive power loop before the compensation.

[2]

400

[3]

300

[4]

Voltage [V]

200
100

[5]

[6]

-100
-200

[7]

-300

[8]

-400
-8

-6

-4

-2

Current [A]

C.I. Budeanu: Puissance reactives et fictives. RGE, [Link], 1928, 762773.

L.S. Czarnecki: What is wrong with the Budeanu concept of reactive and
distortion powers and why is should be abandonded. IEEE Trans.
Instrum. Meas. IM-Vol.36, No.1, March 1988, 834-837.
S. Fryze: Moc czynna, bierna i pozorna w obwodach o przebiegach
odksztalconych pradu i napiecia. Przeglad Elektrotechniczny, nr 7, 1931,
193-203, nr 8. 225-234 oraz nr 22, 1932, 673-676.
S. Fryze: Wybrane zagadnienia teoretycznych podstaw elektrotechniki.
PWN, Warszawa Wroclaw 1966.
M. Ilovici: Definition et Measure de la Puissance et de IEnergie
Reactives, Bull Soc. Franc. Electryciens, LIV, (1925), n. 5, 931-956.
N.L. Kusters, W.J.M. Moore: On the definition of reactive power under
nonsinusoidal conditions. IEEE trans. Pow. Appl. Systems. Vol.99, 1980,
1845-1854.
W. Shepherd, P. Zakikhani: Suggested definition of reactive power for
nonsinusoidal systems. Proc. IEEE Vol.119, No.6, 1973, 704-706.
S. Slabosz: Moc bierna w obwodach jednofazowych z okresowymi
przebiegami, V Konferencja Naukowa Studentow Politechniki
Wroclawskiej, Wroclaw, 2007.

Figure 5. Reactive power loop after the compensation loop area equals zero.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
352

Grid integration of renewable energy sources

P. Kammer, A. Kober
BTU Cottbus
Cottbus, Germany

I. INTRODUCTION
The growing lack of fossil fuels, the increasing demand for
electricity and the harmful effect of carbon dioxide output on
the climate force nations - especially industrialized countries
and their governments - to find new ways of producing the
amount of energy in demand. The integration of alternative
energies to reduce emissions and to conserve available fossil
sources is a known political aim. EU directive 2001/77/EC
requires feeding in renewable energies into electricity grids.
Although the potential of renewables is very high and the
technical conditions to produce electricity are achieved, the
current generated rates are clearly lower. A key problem is the
integration of renewable energies into the existing grid. This
paper analyzes the reasons for this deficit and assesses possible
solutions.
II. RENEWABLE ENERGY SOURCES
There are different options for producing electricity from
renewable energy sources. Consequently, there are several
ways of connecting the gained electricity with the existing grid.

Wind

Hydropower

Solarplants

Preferred sources are wind, hydro, solar, biomass, photovoltaic

cells, bio fuels and geothermic (illustration 1). The electricity
is induced by asynchronous or synchronous generators except
for photovoltaic cells. This operation creates co-current flows
and gets through an inverted rectifier into the power grid.
III. INTEGRATION IN EXISTING GRID / BARRIERS
To understand the problem of why the potential of renewable
energy sources is not exhausted, you have to consider the
actual grid conditions and the resulting barriers. The public
electricity supply in most European countries presents an
extensive central configuration. Grids are designed to transmit
electricity generated by large conventional power plants. An
aggregation occurs by using transformers between the
transmission and the distribution grid. At the end is the
consumer. Illustration 2 demonstrates that the load flow
mainly takes place in one direction from the highest voltage of
the transmission grid with less bulk consumer to the lower ones
in the distribution grid with many small consumers.1

Biomasspower plants

Grid connection

Photovoltaic cells

Bio fuels

Geothermic

Illustration 1: in dependence on BITSCH, R.; GJARDY, G.; WOLDT, T.:

Aspects of large scale RES/DG integration in existing energy
supply systems -considering as example the situation in
Germany, in: International Journal of
Distributed Energy Resources, Volume 2, Number 1, 2006, S.
59-81)

Illustration 2: B. HASCHE; R BARTH; D. J. SWIDER

(Verteilte Erzeugung im deutschen Energiesystem)
centralizedl energy supply and the electricity direction
(source: Barth 2006)

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
353

The interconnection directly or indirectly allocates the

electricity to connected users in a central way. In the
distribution grid the voltage falls in the direction of the current
flow. The decline depends on resistance and inductance in the
cable. These two factors grow with rising cable length. To
provide all consumers with enough voltage, a transformer
slightly increases the voltage at the beginning of a cable.
Energy generation from renewable sources requires an
installation of the plant in locations with a high energy supply,
for example, in areas with a grand wind velocity. Therefore,
the installations are connected at different local points to the
grid. In contrast to large power plants, renewable plants have
less capacity and are integrated in lower grid levels. When
decentralized generators integrate electricity in low-voltage
lines, conditions can change and the power flows in the
direction of the transformer. In this case voltage levels
increase at the end of a line shown in illustration 3. Voltage
rise aggravates in practice if more and more distributed
generators, especially in pastoral areas with mostly weak grids,
are integrated. This barrier of insufficient grid capacity
available for renewable energy is the main problem.2

output of carbon dioxide.3 To avoid voltage rises, the grid has

to be partly extended. The costs of grid reinforcement are often
very high. The benefits of producing energy from renewable
sources are often considered less important than the costs.
Moreover, alternative power developers have highlighted that
it is impossible to determine the available grid capacity so that
they are unable to verify the technical and cost data of the grid
connection presented to them by the grid operator. Furthermore,
Distribution System Operators (DSO) are often linked to
electricity generation companies. It is disputable whether such
a DSO is fully objective towards independent renewable
energy producers when the electricity generation company is
involved in developing alternative energy programs. The
insufficient transparency of grid connection causes long lead
times to obtain grid connection authorization. Polls show that
stakeholders perceptions of grid barriers per renewable energy
source are very high.4

IV. DECENTRALIZED GENERATION / FUTURE

INTEGRATION
The trend for additional energy supply systems (especially by
using renewable sources) in the medium and low voltage level
is highly visible. But to solve all mentioned barriers and to
afford integration of renewable energy sources, many changes
are necessary, including the configuration of an intelligent
distributed electricity system as the most important alteration.
However, there is no clear definition of this concept. In Europe
it is called decentralized generation. Decentralized energy
supply aims at generating the energy where it is required, or
respectively consuming the energy where it is generated.

Illustration 3: voltage curve by integration of renewable energy in low voltage

level
(source: Fraunhofer ISE; wik consu
Studie fr das Bundesministerium fr Wirtschaft und
Technologie (BMWi))

The integration of distributed energy producers seriously

influences the operation of the whole grid and calls for new
requirements of the mains operation. For example, in Germany
the Erneuerbaren-Energien-Gesetz (EEG) requires the
permanent input of available energy from distributed
renewables plants. Thus central large power plants are forced
to work in part load and have additional starts. These actions
have negative effects on materials, efficiency, costs of
generation and lead to additional input of fossil fuels and

It is necessary to supply the available energy in particular

supply-dependent renewable energy to that load which at the
moment of occurrence has the most urgent demand or is being
used for the optimum purpose. Additionally, supply of an area
has to be economically and ecologically optimized according
to criteria which yet have to be specified.5 The character of a
distribution grid will move from the pure allocation of
electricity to the consumer to an orientation of the generation
measured on the demand of the complete system due to the
accession of distributed generation. Illustration 4 shows the
modification of the whole grid. The main difference is the
generation and integration of distributed producers in the
medium and low voltage level and possible input of electricity
in higher sections if the demand falls below the supply.6

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
354

Illustration 4: B. HASCHE; R BARTH; D. J. SWIDER

(Verteilte Erzeugung im deutschen Energiesystem)
decentralized energy supply and the electricity direction

To avoid interferences and overcharge in circuits, storage

facilities such as accumulator systems, capacitors or
compressed-air stores have to be developed and used.
This operation requires that network operators actively
incorporate distributed generators in the grid management.
Energy management can facilitate such interactions between
different electricity levels. Functions like communication,
information, control, metering, services, planning, optimization
and forecast will be of high importance to manage the grid
system error-free.7
The integration of many decentralized generators can be
concentrated in a virtual plant to increase the efficiency.
Such projects contain the optimization of the energy mix of
different renewable energy source characteristics in one area.
This concentration of many small generators allows the supply
of huge quantities of electricity by considering them as one
large power plant. Such an integration of renewable energy
sources not only supplies additional energy but is a part of the
grids capacity. Virtual plants can also act as microgrids and
work independently from the main grid. Such network
expansions which are not grid-controlled do not aggravate
mains power failure but help to hold the grid steady. Projects in
this vein help to increase the total benefits from renewable
generation and integration by smoothing and fullfillment of
planning.8
IV. SUMMARY
Energy supply will pass through a technological change from a
so far generation-dominated, security- and reserve-thinking
centralized grid to a demand-oriented, economically-/

Illustration 5: EU Research 2002

comparison of available grid systems

ecologically-optimized decentralized grid with many

distributed generators, especially using renewable energy
sources. The plants have to be part of the grid system so that
renewable sources do not only produce energy but are actively
incorporated into grid management and capacity.
The allocation of technologies to store the excess electricity
and control the different processes in the grid with adequate
communication ports is one of the basic requirements for a
better integration of alternative energy sources. A future task
will be to find solutions for a better operation mode of large
power plants, especially in part load so that the benefits of
using renewable energy will not be lost because of low
efficiency. If these conditions are achieved, the integration of
alternative energy sources can help to stabilize the current grid
in perturbations and do not aggravate the situation.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
355

REFERENCES
[1]

[2]

[3]
[4]

[5]

EnergieUniversitt Stuttgart Institut fr

Energiewirtschaft (IER) und Rationelle
Energieanwendung; Verteilte Erzeugung im
deutschen Energiesystem
Bernhard Hasche; Rdiger Barth; Derk Jan Swider
wik-Consult FhG Verbund Energie
Potenziale der Informations- und KommunikationsTechnologien zur Optimierung der Energieversorgung
und des Energieverbrauchs
(eEnergy); Bad Honnef, 21. Dezember 2006
Same as in [1]
OPTRES: Assesment and optimisation of renewable
energy support schemes in the European electricty
market (february2007); Intelligent Energy Europe
Lehrstuhl Dezentrale Energiesysteme und
Speichertechnik [Link]. Rainer Bitsch BTU
Cottbus
Aspects of grid integration of decentralized generation
(energy management and LSVPP)

[6]

[7]

[8]

Frauenhofer ISE
Electrical Energy Distribution Networks: Actual
Situation and Perspectives for Distributed Generation
([Link]
ch-jahrgaengen/2003/electrical-energy-distributionnetworks-actual-situation-and-perspectives-fordistributed-generation)
Lehrstuhl Dezentrale Energiesysteme und
Speichertechnik [Link]. Rainer Bitsch BTU
Cottbus
Integration Erneuerbarer Energien in die
Stromversorgung/ Technische Anforderungen an
dezentrale Versorgungsstrukturen in Europa
Dr.-[Link] Schwaegerl, Dipl.-Ing. Anton Heher
OGE-Fachtagung Realitt und Vision der
kologischen Stromversorgung

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
356

Solar potential of the Sahara Desert with an

introduction to solar updraft power plants
Stefanie Fiedermann, Jadranka Halilovic and Torsten Bogacz

Abstract Considering the rising energy demand, it is

necessary to support new sources and possibilities for
energy extraction. Renewable energies play a determining
role. Amongst others, it is inescapable to regard
geographic regions with extreme climatic circumstances.
The desert Sahara has a huge potential of solar energy. Its
dimension and climatic circumstances give cause to inspect
those factors, which would permit solar power generation.
An excellent way to use the premises is e.g. solar updraft
power
plants.
Index Terms Sahara, updraft, Renewable energy,
solar potential, wind
I. INTRODUCTION
As a consequence of the rising worlds demand for electric
energy it is necessary to use alternative possibilities to the
conventional generation of electricity.
The focal point is on the renewable energies. The sun plays a
major role if its potential is used properly. It produces huge
amounts of energy. In one hour it radiates enough energy to
earth to cover the yearly energy demand of the whole world. If
just 0.1 % of the solar energy would be used, the problems of
energy supply could be solved (Figure 1).

The desert Sahara extends over nine million square kilometer

and thus gets the most solar radiation in the world. [4] Due to
the geographical position and the resulting low clouding the
sunshine duration is very high. This leads to a long-time
average of 4000 hours, or eleven hours of sunshine a day in
the deserts heart. This number is shortened to 3285 hours per
year (9 hours/day) on the outside margins.
The sunshine hours of selected sunny meteorological stations
in Germany are well below 2000 hours. [3]
This huge potential of available solar power combined with
the very low population density predestines the Sahara as a
central location for the generation of electricity. The condition
for this is the realization of the transfer of electric energy over
long distances with a minimization of losses.
One option for using the solar power is the solar updraft
power plant. The typical build-up will be described below.
In this power plant the sun heats up the air mass which is
situated below the solar collector. Due to a central chimney
occur differences of pressure. The following equalization of
pressure generates up winds. One or more turbines produce
electric energy as a result of the airflow. [1]
This method of generating electricity was described by Isidoro
Cabanyes for the first time in 1903. The pilot scheme in the
1980s should prove the technical feasibility as well as the
efficiency for a period of years.

II. GLOBAL AND LOCAL FACTORS

Fig. 1: yearly solar radiation on earth

Capabilities for the use of solar power are especially sunny

regions. The generated electric energy, which gets along with
almost no consumption, could be used primary for the own
requirements of this region and later for the external trade by
exporting this electricity. There is no necessity for fossil
burning (coal, oil, gas) moreover the investment grows and a
lot of jobs can be created.
Especially developing countries would profit from work and
energy. Furthermore the generation of updraft energy is
environment-friendly and inexhaustible.

Manuscript received March 20, 2009.

Stefanie Fiedermann is a student of industrial engineering and electrical
engineering at the Brandenburgische Technische Universitt Cottbus,
Germany
(e-mail: [Link]@[Link]).
Jadranka Halilovic is a student of industrial engineering at the
Brandenburgische
Technische
Universitt
Cottbus,
Germany
(e-mail: [Link]@[Link]).
Torsten Bogacz is a student of industrial engineering at the
Brandenburgische
Technische
Universitt
Cottbus,
Germany
(e-mail: [Link]@[Link]).
8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
357

III. MODE OF OPERATION

b)

Chimney

The main task of the chimney is the conversion of the heat

energy into kinetic energy. Furthermore, the chimney is an
important constituent because the performance of solar
updraft power plant depends largely on two factors: on the
one hand the surface of collectors and on the other hand the
height of the chimney.
There are different solutions for the chimney construction
(Figure 3), e.g. out of Ferro concrete or as a steel construction,
which should be taken under consideration of the location.
This is a crucial point, whether it is about the investmentcalculation or construction. The primary factors, which should
take attention, are the wind- and climate-terms, but mainly the
strength and commonness of earthquakes. [1]

The mode of operation of updraft power plants is based on the

three physical doctrines: greenhouse effect, chimney draft and
transformation of kinetic energy.
The build-up of such energy-producing facilities consists
basically of three essential components. These are the solar
collectors, the chimney and the turbines with generator and
gear (Figure 2).

Fig. 2: typical construction of a solar updraft power plant

The solar radiation impacts on the solar collectors and on the

water reservoir which heats up the air mass and the ground
below. The consequence of this is a pressure difference
between the chimney and the heat store.
Through the lesser density of the warm air mass under the
glass roof it rises up in the chimney. The Stack Effect
intensifies the pull of more air mass.
Due to these effects a continuous upwind is ensured for
driving one or more turbines. The gear transmits the mechanic
energy to the generator, which finally transforms it to electric
energy.
One big advantage of solar updraft power plants is the
possibility of a 24/7 operation. This is possible because of the
constant heat emission of the heat stores. These are realized by
water tanks under the glass roof as well as heated ground.
Therefore the heat, saved during the day, can be released in
the night. This opportunity makes the operation possible
during darkness and without fossil burning. [2]

Fig. 3: different kinds of chimney construction

c)

Turbine and generator

By dint of turbines it is possible to detract the mechanical

energy of the airflow.
Several turbines fill the diameter of the chimney and are
driven by the constant airflow in the chimney (updraft).
Through the kinetic energy of the turbines a generator is
provided which then converts the accrued energy into
electricity. [1]

IV. COMPONENTS OF SOLAR UPDRAFT

POWER PLANT
a)

Solar collectors

This basic component of the solar updraft power plant consists

of glass- and foil-cover and uses the greenhouse effect. For
instance, a 200-megawatt disposition would need eight
kilometers diameter of solar collector area and a 1000 metres
tower. The warm air mass, which is the result of the heating
under the glass-roof, flows radial through the collector
towards the base
of theInternational
chimney, where
it rises
[1]
8. EEEIC
Conference
onup.
Environment
and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
358

3
REFERENCES
[1] dos Santos Bernardes, Marco Aurlio (2002): Technische,
konomische und kologische Analyse von
Aufwindkraftwerken
[2] Egger, Daniel (2002): Moderne Solartechnologien und
ihre zuknftigen Perspektiven. ETH Zrich.
[3] Kulzer, Wolfgang (2003): Extremklimate: Zentrale
Sahara
[4] Seefeldt, Katja (2002): Strom aus der Wste:
[Link]

FIGURE REFERENCES
[1] Brtting, Benedikt (2005): Regenerative Energien:

[Link]

[2] dos Santos Bernardes, Marco Aurlio (2002): Technische,

konomische und kologische Analyse von
Aufwindkraftwerken
[3] Schlaich, Jrg (2002): Aufwindkraftwerke:
Fig. 4: solar updraft power plant

[Link]
2_05_03.pdf

[4] Die

Zeit

(2008):

Das

Aufwindkraftwerk,

[Link]

V. CONCLUSION
The solar updraft power plant distinguishes through a simple
concept in connection with a simple construction. The basic
components consisting of collectors, turbines and chimney do
not have to be newly developed, solely an adaptation and
possibly advancement is necessary.
Important assumptions for the operation of such power plant
are a capacious surface for the collectors and long solar
radiation. This can be warranted in areas like the Sahara desert
e.g. another advantage of the Sahara location is the closeness
to Europe. Considering the fact of the high energy demand in
Europe, the EU states could be potential electricity importer.
Therefore, it is certainly fundamental that low-loss energy
transfer is assured.
By achieving the mentioned requirements, the solar updraft
power plant could develop into an environmentally friendly
power generation alternative and become an important
economic factor in an economically underdeveloped region.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
359

Power Quality in distribution power networks with

photovoltaic energy sources
Ricardo Albarracn, Hortensia Amars
Universidad Carlos III de Madrid.
Madrid. Spain
[Link]@[Link], hamaris@[Link]
Abstract- Solar radiation is characterized by short fluctuations
introduced by passing clouds. These solar fluctuations will
produce Voltage and power fluctuations at the PCC (Point of
common coupling). Flicker level should be evaluated by using a
flickermeter according to the standard IEC 61000-4-15. Models of
the solar fluctuation, photovoltaic modules and power converter
are shown in this paper and the flickermeter model is tested
according to the IEC requirements and the CIGRE/CIRED/UIE
test protocol.

Figure. 1. Block diagram of the whole system.

A. Climatology
Index Terms Power quality, Flicker, PV.

I.

INTRODUCTION

Due to the increasing penetration of distributed generation in

electric power systems, power quality is becoming of crucial
importance for the further deployment of renewable generation.
The irregular solar radiation is considered to be one of the
main drawbacks of the large-scale application of photovoltaic
(PV) in distribution networks. Moving clouds can produce fast
and short irradiance fluctuations, which can produce voltage
fluctuations in power networks. This effect is more important
in weak residential and rural grids with high series resistance.
Flicker is defined as the impression of fluctuating brightness
or color, occurring when the frequency of observed variation
lies between a few hertz and the fusion frequency of images
according to the IEEE standard dictionary of electrical and
electronic terms (IEEE standard 100-1977). The flicker level is
dependent on the amplitude of the voltage fluctuation, their
frequency, and the shape of the waveform. All types of voltage
fluctuations may be assessed by direct measurement using a
Flickermeter, which complies with the specification given in
IEC-61000-4-15 [1].
A test protocol is proposed by the CIGRE/CIRED/UIE
voltage quality working group [2] to characterize the
performance of existing flicker meters in the field.
In this paper, voltage fluctuation in power networks with
photovoltaic energy sources will be analyzed and a
flickermeter model will be used for the evaluation of the flicker
assessment under sunny and cloudy situations.
II. DESCRIPTION OF THE MODEL
The scheme used for the implementation of the model is
shown in Fig. 1.

This block represents the effect that weather has on the PV

generation. The outputs of this block are solar radiation G
(W/m2) and temperature T (C). Previous studies modeled this
effect on photovoltaic panels [3], [4]. The effect of sunlight on
a PV and its connection to the network through a converter is
simulated in [5]. The purpose of all these studies is to calculate
the Maximum Power Point Tracking (MPPT) to get the most
use of the solar resource for getting maximum power
production.
In this study, different weather scenarios will be taken into
account. Such as the presence of solar radiation to a greater or
lesser intensity depending on time, day and season or the
presence of darkness caused by obstacles, clouds or other
elements.
B. Photovoltaic module. PV
This block defines the characteristics of solar cells and their
interconnection to the power network.
Some authors have used a model of solar cell consisting of
five parameters [3], [5] that allow obtaining the inherent nonlinear current voltage (I-V) relationship of a typical PV cell. To
obtain the curves we use the features provided by the modules
manufacturers. In this study the model of the solar cell shown
in Fig.2 will be used.
In the presence of solar radiation, the diode (D) produces a
current called Isc (Short-circuit current). This current is
directly proportional to the solar radiation on the cell. Cell
works as a diode. During darkness the solar cell is not an active
device.
A control method for calculating the MPPT point will be
applied. This model will allow obtaining the current Idc and
the voltage Vdc, of the PV in these conditions.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
360

Figure. 2. Equivalent circuit model of a solar cell.

C. Converter DC/AC
This block contains a voltage converter from the PV voltage
Vdc to the Vac that is injected to the power network. For the
connection to the network it is necessary to use a power
transformer at the PCC (Point of Common Coupling). Both are
included in this block.
D. Flickermeter model
The flickermeter model [1] is composed by the following 5
blocks:

Block 1: Input voltage adaptor

This block includes a voltage adapting circuit that scales
the RMS value of the input voltage down to an internal
reference level.

Block 2: Squaring
This block simulates the squaring part of the lamp model.
The lamp model consists of a squaring function and a low
pass filter function.

Block 3: Filtering
a) Demodulator filter
The demodulator filter consist of a first order high pass
filter for suppressing the direct current component and a
low pass filter for suppressing all components equal to or
greater than the fundamental frequency of the carrier
voltage.
b) Weighting filter
The weighting filter simulates the frequency response of a
coiled coil filament gas filled lamp and the human visual
system. The transfer function is of the following type:

(1 + s 2 )
k1s
FW ( s ) = 2
2
s + 2s + 1 (1 + s 3 )(1 + s 4 )
Where s is the Laplace complex variable and the constant
values are given in [1].

Block 4: Variance estimator

This block, called non-linear variance estimator, is
composed by a squaring multiplier and a first order low
pass filter with a time constant of = 300ms. The purpose
of this block is to simulate the storage effect of the human
brain. The time signal at the output of block 4 represents
the instantaneous flicker sensation, Pinst,max.

Block 5: Statistical analysis

Block 5 basically represents a statistical method that
classifies and computes the short term Flicker level Pst.

In this study, the flickermeter model will be used to obtain

flicker level (Pst) at the PCC where both the photovoltaic plant
and the residential area are connected.
III. EXPERIMENTAL RESULTS
The current version of IEC61000-4-15 [1] specifies a
performance test limited to a set of rectangular voltage
fluctuations. The CIGRE/CIRED/UIE Joint Working Group on
Voltage Quality has defined several different test voltage
patterns [2], (seeTable I).
The additional tests with specific voltage fluctuation patterns
are introduced with the goal to standardize the technical
implementation of Flickermeters. In comparison tests it was
shown that commercial Flickermeters connected in parallel to
the same voltage source with arbitrarily selected voltage
fluctuation patterns, yielded substantially different Pst values,
even though all Flickermeter manufacturers claimed that their
products meet the voltage fluctuation tests specified in IEC
61000-4-15 [1].
A. Test #1: Rectangular Voltage Modulation Performance.
For all voltage fluctuations tabulated in Table II the flicker
severity indicator Pst must be within the range of Pst =
1.000(1.00 5%). Test that did not pass the Test Protocol are
green marked.
TABLE I
Summary of Test Protocol

Category
IEC 61000-4-15

No Influence test

Influence test

Complex
Systematic

Test

Excitation Signal

1
2
3
4
5
6
7
8
9
10

Rectangular fluctuation
Rectangular fluctuation
Sinusoidal fluctuation
Frequency variation
High Frequency
Linearity
Single Interharmonic
Harmonic-Interharmonic Pairs
Phase Jump
Interruptions

11

Wood Chipper pattern

Rolling Mill pattern
Arc Furnace pattern

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
361

Output
Pst

Pinst,max
Pinst,max
Pst

Pinst,max
Pst

Pinst,max
Pinst,max
Pst
Pst t
Pst

Rectangular
changes per
minute
1
2
7
39
110
1620

TABLE II

TABLE VI

Results for Test #1

Results for Test #6

Pst results
120-V lamp
230-V lamp
60
Hz
system
system
0.9999
1.0135
0.9758
0.9911
0.9952
0.9957
Test
not
required

4000

Pst results
50

0.9830
1.0170
0.9804
0.9933
0.9923
0.9972
1.2385

120-V
Hz lamp 50
Hz
system
1.0015
1.0137
0.9864
0.9917
1.0010
0.9964
1.2383

230-V
lamp 60
Hz
system
0.9827
1.0155
0.9795
0.9947
0.9927
0.9968
Test not
required

Test not
1.0575
required
Note: 1620 rectangular changes per minute corresponds to a rectangular square
wave modulation frequency of 13.5 Hz.
4800

1.0573

Test not required

TABLE III
Results for Test #2

Pst results

Pst results

Hz

120-V lamp 60
Hz system

230-V lamp 50
Hz system

120-V lamp 50
Hz system

230-V lamp 60
Hz system

0.5
3.5
8.8
18
21.5
22
25
25.5
28
30.5
33+1/3
37
40

0.9960
0.9985
1.0031
0.9993

0.9902
0.9919
0.9933
0.9872
0.9876

0.9958
0.9962
1.0006
0.9894

0.9905
0.9902
0.9948
0.9937

0.9993

1.005

1.0325
1.0548
0.7194

1.0391
1.0530
1.5428

0.9950

0.9874

0.9939

1.0440
0.9956
1.1231

Pst results
120-V lamp 60 Hz 230-V lamp 50 Hz Pst = (1.00 0.05)*()
system
system
0.1
0.1917
1.9912
4.9547
9.9363
19.773

0.1905
1.9852
4.9587
9.9324
19.74

0.2 0.1
2 0.1
5 0.1
10 0.1
20 0.1

B. Test #2: Normalized Response for Rectangular Voltage

Fluctuations.
This test is to verify that the voltage modulation levels, for a
maximum instantaneous value Pinst,max of 1.00 are within a
tolerance of 0.05 of the voltage modulation percentages
given in IEC 61000-4-15. Table III shows the results for Test
#2.
C. Test #3:Normalized Response for Sinusoidal Voltage
Fluctuations.
To perform a further evaluation of flickermeter
implementations, sinusoidal modulation tests points have been
defined in IEC 61000-4-15, Amendment 1 Table 1. Test #3 is
to verify that the voltage modulation levels for a maximum
instantaneous value Pinst max of 1.00 are within a tolerance of
+/- 5% of the voltage modulation percentages given in IEC
61000-4-15,Amendment 1 Table 1. Table IV shows the results
for Test #3.

0.9877

1.0246
0.9990
1.1186

TABLE IV

D. Test #4: Power Frequency Variation.

The purpose of this test is to verify that a steady state
frequency, within a specified tolerance of the nominal
frequency, does not result in light flicker. A test point passes if
the result is Pst of 1.00 0.05. Table V shows the results for
Test #4.

Results for Test #3

Pst results

Pst results

Hz

120-V lamp 60 230-V lamp 50 120-V lamp 50 230-V lamp 60

Hz system
Hz system
Hz system
Hz system

0.5
1.5
8.8
20
25
33+1/3
40

0.9996
0.9999
0.9999

0.9939
0.9934
0.9937

0.9926
1.0518
1.1311

1.0040
1.5310

0.9996
0.9999
1.0000
1.0012

0.5304
0.9934
0.9937
0.9920

1.5427

1.0450
1.1244

TABLE V
Results for Test #4

Power frequency (Hz)

Pst

49
49.5
50.5
51

0.9925
0.9926
0.9956
0.9948

E. Test #6: Linearity.

This test will show whether the measured Pst has a linear
relationship to the amplitude of the voltage fluctuations in the
range 0.2 to [Link] VI shows the results for Test #6.
IV. CONCLUSIONS.
In this paper, a detailed model of the flickermeter according
to the IEC 61000-4-15 has been shown. The flickermeter
model fulfils the requirements defined in the IEC 61000-4-15
standard. Additionally, it has been tested under additional tests
defined in the CIGRE/CIRED/UIE test protocol.
Once the flickermeter model has been validated, it will be
used for evaluating the flicker severity and the voltage
fluctuations produced by photovoltaic energy sources.
Research studies have proved that irregular solar irradiation
caused by cloud movement can produce voltage and power
fluctuation from PV sources. For voltage fluctuation and
flicker assessment is necessary to use a flickermeter such as the
one shown in this paper.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
362

ACKNOWLEDGMENT
This research has been partially supported by the Spanish
Ministry of Education and Science under contracts ENE200628503 and ENE2008-06504-C02-01.
REFERENCES
[1]

IEC std., 61000-4-15, Electromagnetic Compatibility (EMC) Part 4:

Testing and Measurement Techniques Section 15: Flickermeter
Functional and Design Specifications, 2008.

[2]
[3]
[4]
[5]
[6]

Joint Working Group on Power Quality Cigr C4.1.01/CIRED 2 CCO2/

UIE WG2, Test Protocol for IEC Flickermeter used in Power System
Voltage Monitoring, Draft 11. July, 2004.
F. Gonzlez-Longatt, Model of photovoltaic module in MatlabTM, II
CIBELEC, 2005.
Nabil A. Ahmed, Masafumi Miyatake, A novel maximum power point
tracking for photovoltaic applications under partially shaded insolation
conditions, ELSEVIER, 78 (2008) 777-784.
T. DenHerder., Design and simulation of photovoltaic super system using
Simulink, Electrical Engineering Department, California Polytechnic
State University, San Luis Obispo, 2006.
W. Mombauer, Additional requirements to the IEC flickermeter.
Cigre/Cired/UIE. 2009.

8. EEEIC International Conference on Environment and Electrical Engineering, Karpacz, Poland, May 10-13, 2009
363

Author Index
A
Abbasi, Ataollah 220, 308
Abbasi, Hamid Reza 220, 308
Abbaszadeh, Alireza 263
Adamowicz, Marek 57
Ajayi, Adekunle Babatope 151
Al-Awaad, Ahmad-Rami 175
Al-Hasawi, Wael 155, 185
Alvarez, Carlos 1
Amaris, Hortensia 1, 360
Anczykowska, Agata 83
Anil, Aditya 19
B
B, Chitti Babu 107
Bartkiewicz, Stanislaw 83
Belmadani, B. 141
Benghanem, M. 141
Benzaoui, Mohamed Lassad 124
Berkani, Aberrahmane 118
Bogacz, Torsten 357
Bork, Lars 333
Borowski, Gabriel 110
Brose, Nico 253
C
C, Poongothai 107
Cerny, Vaclav 23
Chlodnicki , Zdzislaw 65
Conlon, Michael 91
D
Dbowski, Marcin 230, 249
Danielski, Piotr 249
Daoud, Ahmed A. 127
Dawidowski, Pawel 339
Deshmukh, Sameer 263, 268
Djilali, Kairous Djilali 141
Dolega, Waldemar 133
Domingues, Nuno Soares 226, 287
Doskocz, J. 179
Drobik, Tomasz 260
E
Egorov, Mikhail 57
Estebsari, Abouzar 341
Ewert, Pawe 318
F
Fantik, Josef 95
Farshad, Siamak 204
Fialova, Andrea 23
Fiedermann, Stefanie 357
Frbel, Anke 280
G
Garnitz, Bastian 333
Gaworska-Koniarek, Dominika 54
Ghiasi Varzeghan, Mahdi 114
Gilany, Mahmoud 155, 185
Golkhah, Mohammad 145
Gono, Radomir 45
Grzelak, Dennis 253
H
Halilovic, Jadranka 357
Havel, Petr 49

Heideck, Guenter 103

Heuer, Maik 103
Horacek, Petr 49, 95
Hoseini, Sied Hadi 114
Houdova, Lenka 32
Houdova, Lucie 23
Hoyer, Robert 293
Hrabcik, Miroslav 45
Hui, Kang Kin 200
I
Iwan, A. 179
Iwanski, Grzegorz 65
Izykowski, Jan 300
J
Janecek, Eduard 23, 32
Janecek, Petr 23
Janik, Przemyslaw 1
Jankow, Michal Piotr 321
Jelinek, Libor 32
Jhapte, Rajkumar 164
Jose, Linju 243
K
K B, Mohanty 107
Kaebisch, Mathias 103
Kammer, P. 353
Kepka, Jakub 325
Khambia, Yogesh 272
Khan, Umar Naseem 212, 232
Kitscha, Stefanie 314
Kober, Andreas 353
Koczara, Wlodzimierz 65
Krisch, Thomas 333
Kulinski, R. 190
Kurnatowski, Rafal 218, 247
Kuscheck, Robin 337
L
Labbaf Khaneiki, Majid 27
Leccese, Fabio 8
Leonowicz, Zbigniew 1, 193
Lesan, Saeed 263
Linh, Nguyen Tung 36
Lu, Yan 346
M
Makw, Andrzej 216
Mackowiak, Lukasz 304
Maliga, Grzegorz Pawel 99, 137, 236
Mazniewski, Piotr 300
Mohanta, Dusmanta Kumar 19
Mohd. Ja'afar, Mohammad Nazri 200
Mortazavi, Seyed Saeidollah 196
Mortezapour, Vahid 263
Mysliwiec, Jaroslaw 83
N
Napierala, Artur 73, 77
Nekoui, Mohammad Ali 145
Noack, Martin 337
Novak, Ondrej 95
O
Ozimek, Mariusz 54

P
Palewicz, Marcin 179
Panda, Ajit Kumar 19
Patil, Amar Mohan 99, 137, 236
Patil, Pankaj M. 170
Pawlowski, Maciej Stanislaw 160
Permatasari, Rosyida 200
Pliszczak, Robert 329
Podzemsky, Jiri 41
Pyda, Daniel 296
Pytlarz, Jan Stanislaw 318
R
Radmanesh, Hamid 220, 308
Rahmati, Hojat 145
Rashtchi, Vahid 114
Rezine, Hacene 118
Rezmer, Jacek 1
Rostami, Mehrdad 220, 308, 341
Roy, Diptendu Sinha 19
Ruczewski, Piotr 1, 79
Rusek, Stanislav 45
Rmer, Henry 293
S
Sabosz, Sebastian 349
Saadati, Marjan 196
Sadeghzadeh, Seyyed Mohammad 341
Saffar Shamshirgar, Sahar 145
Sanchez, Ricardo Albarracin 360
Saudhof, Jonathan 293
Sayed, Ahmad Galal 13
Schade, Wolfgang 41
Schlter, Sren 337
Schmeckebier, Jrgen 314
Schwartze, Carsten 333
Schnemann, Lennardt 253
Sek, D. 179
Sikorski, Tomasz 79
Singh, Mahesh 164
Sonawane, Naresh D. 170
Sonawane, Sayali Y. 170
Sonawane, Yogesh Devidas 170
Staszewski, Luke 208
Steckel, Marcel 314
Strek, W. 179
Strnad, Tomas 95
Strzelecki, Ryszard 57
Stuermer, Henryk 61
Styczynski, Zbigniew 103
Sunderland, Keith 91
Swistacz, Bronislaw 183
Szafron, Cezary Pawel 257
Sznitko, Lech 83
Szubzda, B. 190
Szymanski, Dariusz 240
T
Takbiri, Farzad 204
Tivari, Vaibhav 164
V
Verstege, J. F. 87
Vinnikov, Dmitri 57
Vller, Steve 87

W
Waclawek, Zbigniew 1
Wilczynski, Wieslaw 54
Y
Yazdi, Ali Asghar Semsar 27
Yerekar, Sachin Ramrao 283
Z
Ziaja, Jan 73
Zylka, Pawel 69

PRZEDSIBIORSTWO ENERGETYKI CIEPLNEJ

SP. Z O.O. W JELENIEJ GRZE
Municipal Heat Supply Company (PEC) Ltd. in Jelenia Gra is a local leader and a main supplier
of the system heat in the city. The company is also providing services associated with the sales
and consulting in the field of electric energy and specialist repair and maintenance services.
PEC offer is based on the long-term practice and the specialist know-how of employees and
experience acquired in cooperation with customers, allows achieving goals using the most
economic solutions.
Experience and knowledge:
- qualified professionals
- the production, distribution and sale of the thermal energy, electric and of technological
steam
- providing special services
Technology:
- preinsulated piping
- weather automation
- radio meter reading
Ecology:
- cogeneration
- limiting of greenhouse gas emission
- reducing network losses
Local activities:
- Sponsor of the education and culture
in the region
- Fair Play Company

Future:
- growth in production of the electric energy
- investments in infrastructure
- development of the renewable sources of energy
PEC Jelenia Gra safe and trusted partner in the field of heating technologies.
More informations on our webpage: [Link]. Welcome.

[Link].p

Founded in 1949, the Institute of Power Systems Automation Ltd. is an independent

research and development organisation specialising in automated control and
monitoring syastems for the electricity industry.

51-618 Wrocaw; ul.Wystawowa1;

POLAND
Tel. +48-071-348 42 21
Fax +48-071-348 21 83

ISBN

978 3 940471 10 - 9

2009, Technical University of Cottbus, Informations-, Kommunikations- und Medienzentrum