ECEN 667

Power System Stability

Lecture 20: SMIB Eigenvalues,
Measurement-Based Modal Analysis

Prof. Tom Overbye

Dept. of Electrical and Computer Engineering
Texas A&M University
overbye@[Link]
 Announcements

• Read Chapter 8
• Homework 6 is due on November 11
• There is a 2019 NERC document on oscillations at
[Link]/comm/PC/SMSResourcesDocuments/I
nterconnection_Oscillation_Analysis.pdf

1
 Single Machine Infinite Bus

• A quite useful analysis technique is to consider the

small signal stability associated with a single generator
connected to the rest of the system through an
equivalent transmission line
• Driving point impedance looking into the system is
used to calculate the equivalent line's impedance
– The Zii value can be calculated quite quickly using sparse
vector methods
• Rest of the system is assumed to be an infinite bus with
its voltage set to match the generator's real and reactive
power injection and voltage
2
 Small SMIB Example

• As a small example, consider the 4 bus system shown

below, in which bus 2 really is an infinite bus
Bus 1 Bus 2
X=0.2
GENCLS Bus 4 Infinite Bus
X=0.1
slack

11.59 Deg
Bus 3
X=0.1 X=0.2
1.0946 pu

6.59 Deg 4.46 Deg 0.00 Deg

1.046 pu 1.029 pu 1.000 pu

• To get the SMIB for bus 4, first calculate Z44

 −25 0 10 
10
Z44 is Zth in parallel
 0 1 0 0  
Ybus = j → Z 44 = j0.1269 with jX'd,4 (which is
 10 0 −15 0  j0.3) so Zth is j0.22
 
 10 0 0 −13.33 
 Small SMIB Example

• The infinite bus voltage is then calculated so as to

match the bus i terminal voltage and current
Vinf = Vi − Z i I i
While this was demonstrated
*
 Pi + jQi  on an extremely small system
where   = Ii for clarity, the approach works
 Vi 
the same for any size system
• In the example we have
* *
 P4 + jQ4   1 + j0.572 
  =  = 1 − j0.328
 V4   1.072 + j0.220 
Vinf = ( 1.072 + j0.220 ) − ( j0.22) ( 1 − j0.328 )
Vinf = 1.0
4
 Calculating the A Matrix

• The SMIB model A matrix can then be calculated either

analytically or numerically
– The equivalent line's impedance can be embedded in the
generator model so the infinite bus looks like the "terminal"
• This matrix is calculated in PowerWorld by selecting
Transient Stability, SMIB Eigenvalues
– Select Run SMIB to perform an SMIB analysis for all the
generators in a case
– Right click on a generator on the SMIB form and select Show
SMIB to see the Generator SMIB Eigenvalue Dialog
– These two bus equivalent networks can also be saved, which
can be quite useful for understanding the behavior of individual
generators 5
 Example: Bus 4 SMIB Dialog

• On the SMIB dialog, the General Information tab

shows information about the two bus equivalent

PowerWorld case B4_SMIB 6

 Example: Bus 4 SMIB Dialog

• On the SMIB dialog, the A Matrix tab shows the Asys

matrix for the SMIB generator

• In this example A21 is showing

4 , pu 1  −PE ,4   1  −1  
=   = −   ( −1.2812 cos ( 23.94 ) ) 
 4 2H 4   4   6    0.3 + 0.22  
= −0.3753
7
 Example: Bus 4 with GENROU

• The eigenvalues can be calculated for any set of

generator models
• This example replaces the bus 4 generator classical
machine with a GENROU model
– There are now six eigenvalues, with the dominate response
coming from the electro-mechanical mode with a frequency of
1.84 Hz, and damping of 6.9%

PowerWorld case B4_GENROU_Sat_SMIB 8

 Example: Bus 4 with
GENROU Model and Exciter
• Adding an relatively slow EXST1 exciter adds
additional states (with KA=200, TA=0.2)
– As the initial reactive power output of the generator is
decreased, the system becomes unstable (below example is
with a generator reactive power output of 0 Mvar)

PowerWorld case B4_GENROU_Sat_SMIB_QZero 9

 Example: Bus 4 with
GENROU Model and Exciter
• The below image shows the system response to a brief
bus 4 self-clearing fault
90

88

86

84

82

80

78

76

74

72

70

68

66

64

62

60

0 1 2 3 4 5 6 7 8 9 10

Rotor Angle_Gen Bus 4 #1

b
c
d
e
f
g

10
 Example: Bus 4 with
GENROU Model and Exciter
• The remainder of the Eigenvalues page shows the
participation factors for the various states in the modes

11
 SMIB Eigenvalues for TSGC_2000
Case
• All the SMIB eigenvalues can be calculated quickly
even for relatively large grids

12
 Saving a Two Bus Equivalent

• PowerWorld makes it easy to save a two bus

equivalent from the SMIB Eigenvalues page
– Right-click and select Save Two Bus Equivalent
• As the name implies, the two bus equivalent is the
generator connected to an infinite bus through its
driving point impedance
• Two bus equivalents provide a convenient way to track
down at least some causes of instability issues

13
 Small Signal Analysis and
Measurement-Based Modal Analysis
• Small signal analysis has been used for decades to
determine power system frequency response
– It is a model-based approach that considers the properties of a
power system, linearized about an operating point
• Measurement-based modal analysis determines the
observed dynamic properties of a system
– Input can either be measurements from devices (such as
PMUs) or dynamic simulation results
– The same approach can be used regardless of the
measurement source
• Focus in this section is on the measurement-based
approach
14
 Ring-down Modal Analysis
• Ring-down analysis seeks to determine the frequency
and damping of key power system modes following
some disturbance
• There are several different techniques, with the Prony
approach the oldest (from 1795); introduced into
power in 1990 by Hauer, Demeure and Scharf
• Regardless of technique, the goal is to represent the
response of a sampled signal as a set of exponentially
damped sinusoidals (modes)
q
− i
y (t ) =  Ai e cos (i t + i )
 it Damping (%) =  100
i =1
 +
i
2
i
2

15
 Where We Are Going:
Extracting the Modes from Signals
• The goal is to gain information about the electric
grid by extracting modal information from its signals
– The frequency and damping of the modes is key
• The premise is we’ll be able to reproduce a complex
signal, over a period of time, as a set a of sinusoidal
modes
– We’ll also allow for linear
detrending
0.1𝑡 + cos 2𝜋2𝑡

16
 Example: The Summation of two
damped exponentials
• This example
was created by
going from the
modes to
a signal
• We’ll be going
in the opposite
direction (i.e.,
from a
measured
signal to the
modes) 17
 Some Reasonable Expectations

• Verifiable to show how well the modes match the

original signal(s)
– We’ll show this
• Flexible to handle between one and many signals
– We’ll go up to simultaneously considering 40,000 signals
• Fast
– What is presented will be, with a discussion of the
computational scaling
• Easy to use
– This is software implementation specific; results shown here
were done using the modal analysis tool integrated into
PowerWorld Simulator (version 22) 18
 Example: One Signal
This could be any signal; image shows the result of the
original signal (blue) and the reproduced signal (red)

19
 Verification:
Linear Trend Line Only

20
 Verification:
Linear Trend Line + One Mode

21
 Verification:
Linear Trend Line + Two Modes

22
 Verification:
Linear Trend Line + Three Modes

23
 Verification:
Linear Trend Line + Four Modes

24
 Verification:
Linear Trend Line + Five Modes

It is hard to tell a difference

on this one, illustrating that
modes manifest differently in
different signals

25
 A Larger Example We’ll Finish With
Applying the developed techniques to the response of all 43,400
substation frequencies from an 110,000 bus electric grid(20 million
plus values)

26
Spatial Visualization of Frequency

27
 Measurement-Based Modal Analysis

• There are a number of different approaches

• The idea of all techniques is to approximate a signal,
yorg(t), by the sum of other, simpler signals (basis
functions)
– Basis functions are usually exponentials, with linear and
quadratic functions used to detrend the signal
– Properties of the original signal can be quantified from basis
function properties
• Examples are frequency and damping
– Signal is considered over time with t=0 as the start
• Approaches sample the original signal yorg(t)
28
 Measurement-Based Modal Analysis

• Vector y consists of m uniformly sampled points

from yorg(t) at a sampling value of T, starting
with t=0, with values yj for j=1…m
– Times are then tj= (j-1)T
– At each time point j, the approximation of yj is
n
yˆ j (t j , α ) =  bii (t j , α )
i =1

where α is a vector with the real and imaginary eigenvalue components,

i t j
with i (t j , α ) = e for  i corresponding to a real eigenvalue, and
it j it j
i (t j , α ) = e cos( i+1t j ) and i +1 (α) = e sin( i+1t j )
for a complex eigenvector value
29
 Measurement-Based Modal Analysis

• Error (residual) value at each point j is

rj (t j , α) = y j − yˆ j (t j , α)
• The closeness of the fit can be quantified using the
Euclidean norm of the residuals
1 m 1

2
( y j − yˆ j (t j , α)) = r (α) 2
2

2 j =1 2
• Hence we need to determine  and b
n
yˆ j (t j , α ) =  bii (t j , α )
i =1

30
 Sampling Rate and Aliasing

• The Nyquist-Shannon sampling theory requires

sampling at twice the highest desired frequency
– For example, to see a 5 Hz frequency we need to sample the
signal at a rate of at least 10 Hz
• Sampling shifts the frequency spectrum by 1/T (where
T is the sample time), which causes frequency overlap
• This is known as aliasing, which
can cause a high frequency
signal to appear to be a lower
frequency signal
– Aliasing can be reduced by fast sampling and/or low
pass filters
Image: [Link]/wikipedia/commons/thumb/2/28/[Link]/[Link] 31
 One Solution Approach:
The Matrix Pencil Method
• There are several algorithms for finding the modes.
We’ll use the Matrix Pencil Method
– This is a newer technique for determining modes from noisy
signals (from about 1990, introduced to power system
problems in 2005); it is an alternative to the Prony Method
– The Matrix Pencil Method is useful when there is signal noise
• Given m samples, with L=m/2, the first step is to form the
Hankel Matrix, Y such that
 y1 y2 y L +1 
 y y3 yL + 2 
Y= 2
This not a sparse matrix  
 
 ym − L ym − L + 1 ym 

Reference: A. Singh and M. Crow, "The Matrix Pencil for Power System Modal Extraction," IEEE Transactions on Power
Systems, vol. 20, no. 1, pp. 501-502, Institute of Electrical and Electronics Engineers (IEEE), Feb 2005. 32
 Algorithm Details, cont.
• Then calculate Y’s singular values
using an economy singular value decomposition
(SVD) The computational
Y = UΣV T
complexity increases
• The ratio of each singular value with the cube of the
is then compared to the largest number of
measurements!
singular value c; retain the ones
with a ratio > than a threshold
– This determines the modal order, M This threshold
– Assuming V is ordered by singular is a value that
values (highest to lowest), let Vp be can be changed;
then matrix with the first M columns of V decrease it to
get more modes.
33
 Aside: The Matrix Singular Value
Decomposition (SVD)
• The SVD is a factorization of a matrix that generalizes the
eigendecomposition to any m by n matrix to produce
The original concept is more than
Y = UΣV T
100 years old, but has found lots of
recent applications
where S is a diagonal matrix of the singular values
• The singular values are non-negative, real numbers that can
be used to indicate the major components of a matrix (the
gist is they provide a way to decrease the rank of a matrix)

34
 Aside: SVD Image Compression
Example

Images can be
represented with
matrices. When
an SVD is applied
and only the
largest singular
values are retained
the image is
compressed.

Image Source: [Link]/~goller/F15_M2270/BradyMathews_SVDImage.pdf

35
Matrix Pencil Algorithm Details, cont.
• Then form the matrices V1 and V2 such that
– V1 is the matrix consisting of all but the last row of Vp
– V2 is the matrix consisting of all but the first row of Vp
• Discrete-time poles are found as the generalized
eigenvalues of the pair (V2TV1, V1TV1) = (A,B)
• These eigenvalues are the
If B is nonsingular (the
discrete-time poles, zi with the
situation here) then the
modal eigenvalues then generalized eigenvalues
are the eigenvalues of
The log of a complex -1
ln( zi ) B A
i = number z=r is
T ln(r) + j

36
 Matrix Pencil Method with Many
Signals
• The Matrix Pencil approach can be used with one
signal or with multiple signals
• Multiple signals are handled by forming a Yk matrix
for each signal k using the measurements for that
signal and then combining the matrices
 y1,k y2 ,k yL +1,k 
 y y3 , k yL + 2 ,k  The required
Yk = 
2 ,k

  computation
  scales linearly
 ym − L ,k ym − L +1,k ym ,k 
with the number
 Y1 
of signals
Y =  
 YN 
37
 Matrix Pencil Method with Many
Signals
• However, when dealing with many signals, usually the
signals are somewhat correlated, so vary few of the
signals are actually need to be included to determine
the desired modes
• Ultimately we are finding
n
y j (t j , α ) =  bii (t j , α )
i =1

• The  is common to all the signals (i.e., the system

modes) while the b vector is signal specific (i.e., how
the modes manifest in that signal)

38
 Quickly Determining the b Vectors

• A key insight is from an approach known as the

Variable Projection Method (from Borden, 2013) that
for any signal k
y k = Φ(α )b k
And then the residual is minimized by selecting b k = Φ(α ) + y k
where Φ(α ) is the m by n matrix with values
it j Where m is the
 ji (α ) = e if  i corresponds to a real eigenvalue,
number of
t t
and  ji (α ) = e cos( i +1t j ) and  ji +1 (α ) = e sin( i +1t j ) measurements
i j i j

for a complex eigenvalue; t j = ( j − 1) T and n is the

Finally, Φ(α ) + is the pseudoinverse of Φ(α)
number of modes

A. Borden, B.C. Lesieutre, J. Gronquist, "Power System Modal Analysis Tool Developed for Industry Use," Proc. 2013
North American Power Symposium, Manhattan, KS, Sept. 2013
39