Control Problem
System Description
A single machine infinite bus system is shown in
figure below.
vt
i
v
Z R jX
G
P+jQ
YL G jB
Single machine infinite bus system with local load
Symbols
Pm
P (Pe)
R
X
Z
YL
G
Mechanical input power
Electrical output power of the generator
Transmission line resistance
Transmission line reactance
Transmission line impedance
Load admittance
Load conductance
B
i
id, iq
Vt
v d , vq
Tdo
xd
xd
xq
Eq '
Efd
Vref
v (vb or 0)
M
D
KA
TA
uPSS
iL
b
f
Load susceptance
Generator current
d- and q-axis generator current
Terminal voltage
d- and q-axis terminal voltage
Field time constant
d-axis reactance
d-axis transient reactances
Generator q-axis reactance
Generator internal voltage
Field voltage
Reference voltage
Infinite bus voltage
Machine inertia coefficient
Machine damping coefficient
Gain of the excitation system
Time constant of the excitation system
PSS control signal
Load current
Generator rotor angle
Generator rotor speed
Synchronous speed
Frequency (60 Hz)
Nonlinear Model
Let
vt = vd + jvq
i = id + jiq
v = vd + jvq = |v| (sin + jcos )
From the system shown,
vt v
i Yvt
Z
or,
Zi (1 ZY )vt v
For convenience, the following constants and
parameters are introduced.
1 ZY C1 jC2
Thus,
C1 1 RG XB
C2 RB XG
Separating the last equation into real and imaginary
parts, the results can be written in matrix form as
R
X
X id C1
i
R q C2
C2 vd
sin
v v
C1 q
cos
(1)
The state equations of the nonlinear model may be
written
b ( 1)
(Pm Pe )
M
'
q
1
e ' E FD eq' (x d x d' )i d
T do
1
EFD K A (vref vt uPSS ) EFD
TA
where
Pe v d i d v q i q
v d x qiq
(2)
v q eq' x d' i d
(3)
Substitute from (2), (3) into (1)
id Yd ' v
i Y eq 2
Ze
q q
where
R1 R C2 xd'
R2
X
2
X1 sin
R1 cos
(4)
R2 R C2 xq
X1 X C1 xq
X 2 X C1 xd'
Ze2 R1R2 X1 X 2
C1 X 1 C2 R2
Yd
Z e2
C1R1 C2 X 2
Yq
Z e2
Linearized Model
Initial Condition Calculations
The first step to develop the linearized model of the
power system for low frequency oscillation studies is
to calculate the initial conditions such as initial
currents, initial voltages, initial rotor angles etc.
The initial conditions are calculated if Pe0, Qe0, and
|vt0| are given.
The 2nd subscript 0 is to identify the initial values.
The calculations of the initial conditions are as
follows.
Pe0, Qe0, and |vt0| are given.
Then;
Pm =Pe
vd
Pe vt
2 1/ 2
P Q v / x
e
q
2
e
2
t
vq vt2 vd2
iq vd / xq
id
Pe iq vq
vd
eq' vq xd' id
The initial values of v and can be determined from
the matrix equation above as
vd v sin C1vd C2vq Rid Xiq
vq v cos C2vd C1vq Xid Riq
Consequently,
tan 1 (vd / vq )
v v2 d v2 q
EFD eq' ( xd xd' )id
Vref
EFD
vt
KA
Example
A 60-Hz single machine
infinite bus system has the data G
in pu except M and time
constants are in seconds as
follows.
Generator
M=9.26 Tdo=7.76
xd=0.973 xd=0.19
Exciter
KA=50
TA=0.05
Line
R=0.034 X=0.997
Load
G=0.249 B=0.262
Initial state Pe0=1.0 Qe0=0.015
U
PB
PB
v
Z R jX
YL G jB
D=0
xq=0.55
B
vt
vt0=1.05
B
Solution
Calculating the initial conditions gives
vd0=0.4659
vq0=0.9410
id0=0.4354
iq0=0.8471
eq0=1.024
v=1.051
0=68.01
U
PB
Develop a computer program to calculate the initial
conditions.
LINEARIZED MODEL
In state-space form
0
K1
M
K
'4
eq' T do
E fd K A K 5
T A
K
2
M
1
K 3T do'
0
0
Derivation of K1-K6
Define;
R1 R C2 xd'
R2 R C2 xq
X1 X C1 xq
X 2 X C1 xd'
Ze2 R1R2 X1 X 2
C X C R
Yd 1 1 2 2 2
Ze
KAK6
TA
0
0
1 ' 0
e
T do' q K A
E fd
1
T A
T A
u PSS
C1R1 C2 X 2
Yq
Z e2
v
Fd 2 X 1 sin 0 R2 cos 0
Ze
v
Fq 2 R1 sin 0 X 2 cos 0
Ze
Then;
Fq ( xq xd' )iq 0
Yq eq' 0 ( xq xd' )id 0
K1 0 Fd
K i Y
2 q0 d
1
K3
1 ( xd xd' )Yd
K4 ( xd xd' ) Fd
K5 0 Fd
K v v Y
6 q0 t 0 d
Fq xd' vq 0 vt 0
Yq xq vd 0 vt 0
For the above example:
Calculating the K1K6 constants gives
K1=0.5441
K2=1.2067
K3=0.6584
B
K4=0.6981
B
K5= 0.0955
B
K6=0.8159
B
Extend the computer program to calculate the
linearization constants K1-K6 and form the matrices
A and B. Calculate the eigenvalues of the system.
DESIGN OF LEADLAG
CONTROLLER
U
If there is a PSS with the following simple structure
sT w
1 sT w
x5
1 sT1
KC
1 sT2 uPSS
Here, two more states have to be added as follows.
(1 sTw ) x5 sTw
(1 sT2 )uPSS KC (1 sT1 ) x5
In matrix form
K1
K2 ' 1
1
D
x5 x5
eq x5
Tw
M
M
M
Tw
KC
KCT1 1
uPSS
x5
x5 uPSS
T2
T2
T2
i.e.,
K1
K2 ' 1
D
x5
eq x5
M
M
M
Tw
uPSS
KC K1T1
K DT
K KT
K
T
1
C 1 C 2 1 eq' C (1 1 ) x5 uPSS
MT2
MT2
MT2
T2
Tw
T2
Therefore, the system equation with PSS can be
written as
X AX BuPSS
or
Z AC Z
where
Z [ , , eq' , EFD , x5 , uPSS ]T
AC is the controlled or closed-loop system matrix. The
eigenvalues of AC are the closed loop-eigenvalues.
Collecting all equations AC can be written as
0
K1
M
K
'4
T do
AC K A K 5
TA
K
1
M
K KT
C 1 1
MT 2
D
M
0
0
0
K
2
M
1
K 3T do'
D
M
K C DT1
MT 2
KAK6
TA
0
1
T do'
1
TA
K2
M
K C K 2T1
MT 2
0
0
0
0
KA
0
TA
0
Tw
KC
T1
1
(1 )
T2
Tw
T 2
0
Comparing the eigenvalues of A and AC, the
effectiveness of PSS can be assessed.
The effectiveness of PSS can be also assessed through
time-domain simulation by exposing the system to a
disturbance and evaluate the system response.
B
For the above example:
If PSS is given as
uPSS
3s 1 0.6851s
7.09
.
1 3s 1 0.1s
To compare the eigenvalues with and without PSS,
the matrices A and AC are formed. The results are
B
System mode
Mechanical
Electrical
Control
Without PSS
With PSS
0.295 j 4.96
-1.127 j 4.33
-10.393 j 3.284 -4.618 j 7.483
--------------0.3439, -15.726
It is clear that n without PSS is negative as
0.295
0.06
2
2
2
2
0.295 4.96
B
However, with PSS
1.127
1.127 4.33
2
0.252
Extend the program to form the matrices A and AC
and calculate the eigenvalues of the open-loop and
the closed-loop systems.