Design of Robust Power System Stabilizer

for Interconnected System

Praveen Nayak Bhukya1
Abstract Electrical Power System is one of the most complex real time operating systems. It is probably one of the best examples of a large
interconnected non-linear system of varying nature. This system changes it's state time to time with change in load, transmission and
generation conditions. Such changes in the system causes small signal oscillations, the frequency range of such oscillations is about 0.2 to 3
Hz, if they are persist for long duration they may limiting the power transfer capability of the system. For the stable and secure operation of the
system adequate damping for small signal oscillations is required. A auxiliary controller can provide adequate damping for the system called
power system stabilizer (PSS).
Fixed parameter controllers are most commonly used controller for this applications. This is a conventional method of designing
power system stabilizer. It gives the better performance for single operating condition only. In this paper a technique for designing a robust
power system stabilizer using

control theory has been discussed. These stabilizers gives the comparable performance for a wide

range of operating condition. In case of interconnected system conventional method of designing power system stabilizer is extremely
complex, and there is need of system information external to the plant. Modified Heffron-Phillip's model provide solution for this problem, it
allow designing power system stabilizers without using system information external to the plant. In this paper modified Heffron-phillip's model
has used for designing stabilizer for multi-machine system. PSS at each machine can be synthesized using information available at local
buses only.

Index Terms Power system stabilizer (PSS) , small-signal stability, Heffron-Phillips model (HP) , Single Machine Infinite Bus (SMIB).

INTRODUCTION

One of the major problems in electrical power system

operation is related to the small-signal oscillatory
instability caused by insufficient natural damping in
the system. The most effective way of countering this
instability is to use auxiliary controllers called power
system stabilizers (PSS), to produce additional
damping in the system [1], [2]. The concept of PSS and
their tuning procedures are well explored in [1][4].
The xed gain stabilizers perform reasonably well if
they have been tuned properly [5]. Though these
stabilizers have simple robust structures, tuning them
either by computer simulation modeling [2], [4] or by
actual eld tests [3] is an involved process which
requires considerable expertise and also a knowledge of
system parameters external to the generating station.
These parameters may not be readily available and may
vary during normal operation of the power system.
Even in the case of single machine innite bus model,
estimates of equivalent line impedance and the voltage
at the external bus are required. The PSS design also
requires information of the rotor angle measured
with respect to an external bus. These parameters
cannot be measured directly and need to be estimated
based on reduced order models of the rest of the
system connected to the generator. If the available

information for the rest of the system is inaccurate, the

conventionally designed PSS results in poor system
performance.
A coordinated PSS design methodology based
on damping torques approach for a wide range of
operating conditions has been described in [6][8]. This
method uses P-Vr characteristics obtained by disabling
the shaft dynamics of all the machines. However, this
formulation is not suitable for very large systems. In [9]
a thorough analysis of the frequency responses of
generator electrical torques is performed. It is shown
that the frequency responses between AVR input and
the resultant electrical torque at the rotor shaft has two
components. One component is dependent on the
associated generator as well as on the network
admittance matrix augmented with the generator
admittances and the other component depends only on
the associated generator. The diagonal dominance
property of the admittance matrix makes the rst
component less affected by the generators external to
the generator under consideration. It means that the
required dynamic information for the PSS design may
be contained mostly within the generating plant. The
tuning guidelines of [2] recommend the PSS tuning for
a strongest system possible with full loading for speed
and power input signals due to the occurrence of
maximum phase lag under these conditions. From the
above discussions it follows that even if the dynamics

of the external generators are ignored as a rst

approximation and the stabilizer is designed to provide
maximum phase lead around the local mode of
oscillations, the PSS will still have sufficient lead at the
inter area modes which are largely inuenced by the
dynamics of the external generators. Such PSS will be
able to damp out both local mode and inter area modes
of oscillation. The present design is based on this
presumption.
The method proposed for the PSS design in
the present paper is also based on the conventional
design technique as described in [2] and [4]. However,
as opposed to a conventional stabilizer design, the
system dynamics are linearized by taking the
secondary bus voltage of the step-up transformer (high
voltage bus) [10] as reference instead of the innite bus
[11]. The GEP(s) phase characteristics obtained from
this model can be treated as that of a strong system
having virtually transformer reactance as external
reactance with full loading. This model is almost
similar to the Heffron-Phillips (HP) model; however
the model parameters are independent of the external
system information. This facilitates one to use this
model for any machine in the multi-machine
environment. Following conventional compensator
design techniques based on root-locus and bode plots
[12], PSSs
are designed independently for each
machine. All PSS design parameters are thus calculated
from local measurements,

two methods which are based on the complete system

information.

MODELING OF POWER SYSTEM

For small-signal stability analysis, dynamic modeling is

required for the major components of the power
system. It includes the synchronous generator,
excitation system, automatic voltage regulator (AVR),
etc. Different types of models have been reported in the
literature depending upon their specic application.
The model shown in Fig. 1 is used to obtain the
linearized dynamic model [14] (Heffron-Phillips or Kconstant model). IEEE Model 1.0 is used to model the
synchronous generator [12] with a high gain, low time
constant static exciter.

Fig. 2: Single Machine Infinite Bus System

Fig. 1: Single Machine Connected to External Network

i.e., voltage and power measurements at the high

voltage bus and there is no need to estimate or compute
the values of equivalent external impedance, bus
voltage and rotor angle.
The performance of the designed stabilizers is
analyzed for two widely used IEEE test systems, 3
generator 9 bus system and 10 generator 39 bus system.
The performance is evaluated under various operating
conditions and compared with the performance of PSS
designed by PVr characteristics and the method of
residues [13]. The stabilizers based on the proposed
design technique have shown better damping
characteristics under heavy and nominal loading
conditions and more or less similar performance under
light loading conditions, when compared to the other

Fig. 3: linearized Model of Single Machine Infinite Bus System

Generator Mechanical and Electrical Torque Equations:

s = w B S m
(1)

1
[ T m T e D S m ]
2H
'
T e =K 1 + K 2 E q
Sm =

(3)
Expression for Heffron -Phillip's constant

K 2 is given
E E cos 0 X q X 'd
K 1= b q 0
+
Eb sin 0
X q+ X t
X e + X 'd
(4)

(2)

K 1 and

K 2=

Xq+ Xe
X e + X 'd

iq 0

(5)

q-axis Flux Linkage:

E'q =

1
[ K 3 ( E fd K 4 )E 'q
K 3 T 'do

(6)

K3

Expression for Heffron-Phillip's constant

K4

and

is given

K3 =
(7)

K4 =

X e + X 'd
Xd+ Xe
X d X 'd
Eb sin o
X e + X 'd

Fig. 4: Phaser diagram of the system shown in Fig. 1

Excitation System:
Representation of Linearized equations
excitation system can be expressed as

for

the

V t = K 5 + K 6 E 'q
1
TA

( X e + X 'd ) V t 0
K6 =

(11)

V q0
X e + X 'd V t 0
(12)

Consider a single generator connected to the external

system through a power transformer as shown in Fig. 4
The rotor angle with respect to the voltage V s s

s . The expressions for

'
i d and i 'q can be derived
diagram.

tan 1

=
,

'
q

from the phaser

Ia

cos p

Ia

Ps ( X t + X q ) Qs (R a+ R t )
P s ( R a + Rt ) +Q s ( X t + X q ) +V 2s

(16)
Where

Qs

Ps and Qs can be given by

Ps = V s
Ia
cos p
Ia
sin p
= Vs

and

In rare cases, under leading power factor operations

Ps ( R a + R t ) +
(17)
0

Xe

of the high voltage bus is dened as

( X q + X t ) cos p
sin p (13)

(15)

K A ( V ref + V pss V t ) Ef d ] (10)

K 5 and
Expression for Heffron-Phillip's constant
K 6 is given
X q V do Eb cos 0
K5
=

( X q+ Xe )V t 0
X 'd V q 0 Eb sin 0

OAB in Fig. 4 we get

I a ( R a + Rt
sin p (14)
1 AB
= tan
OA

(9)

Ef d

I
AB =
a
R
+
R
(
a
t )
OA = V s
Xq+ Xt )
From the

(8)

And

Qs ( X t + X q )

Vs

is given by

| s obtained (eq 16)|

(18)

The stator algebraic equations are given by

E'q + X 'd i d
X q iq

Ra i q = V q

Ra i d = V d
(19)

From stator algebraic (19) one can get the following

E'q
( X t + X 'd )
=
Xt

equation for
'

Eq

Vt

Xq
V s sin s
( Xt+ Xq)

X 'd
V cos s
Xt s

(20)
The expressions for

id

and

iq

and are as follows

i d = B E'q

s +
)
cos

Y V s
(21)

iq

= G E

'
q

Y Vs
(22)

imaginary parts of the above equation . These

equations are valid even for the multi-machine system.

sin( s + )

(26)

Where

B
= tan 1
2
G
Y eq =

with Y =

V q=R t iq X t i d +V s cos s
V d =Rt i d X t i q +V s sin s
Equating stator algebraic equation (19) and modied
stator algebraic equation (26) and rearranging we can
get

|Y eq|

1
=G+ j B
( ( Ra + Rt ) + j ( X t + X q ) )

( X 'd + X t ) i dR t iq =V s cos s E'q

( X q + X t ) i qRt i d =V s sin s

Modified Heffron-Phillip's Model

(28)

Modied Heffron-Phillip's model is suggested in [7].

The standard linear model of SMIB known as HeffronPhillip's model (also called as K-constant model) can be
obtained by linearizing the system equations around an
operating condition. The synchronous machine can be
interfaced with the external network by converting
machine equations in Park's reference frame to
synchronously rotating Kron's reference frame. The
equations are given below for a SMIB system.

V Q + j V D=( V q + j V d ) e j
j
= ( i q + j i d ) ( Re + j X e ) e + Eb 0

These equations can be written as

[]

By using
Kirchhoffs Voltage law between the
generator terminal and the innite bus. The subscripts
q and d refers to the q and d-axis, respectively, in Parks
reference frame Q and D refers to the q and d-axis,
respectively, in Krons reference frame. Similar equation
can be written between transformer bus and the
generator terminal voltage and is given below. This is
the only modication suggested in this paper to make
the PSS design independent of the external system
parameters

( V q + j V d )=( iq + ji d )( R t + j X t ) +V s s e j

( s + s )

(X

V s cos sE'q
V s sin s

+Xt)

]
9)

X=

( X q + X t ) ( X 'd + X t ) + R 2t

The terminal voltage of the machine is given by

V t = ( V 2d +V 2q )
(30)

Linearizing (29) around an operating condition using

rst order Taylor series approximation and upon
simplication one can obtain
'

i d =C1 s +C 2 E q +CV 1 V s
i q=C 3 s +C 4 E'q +C V 2 V s
(31)
Where

(2
4)
by

Rt
'
d

(2

3)

i d = 1 ( X q + X t )
X
iq
Rt

Where

(2

Replacing

(27)

in the above equation

gives

( V q + jV d )=( i q + ji d )( R t + j X t ) +V s s
(2
5)
We can get the modied stator algebraic equations
referred to the transformer bus by equating the real and

1
R V cos s0 ( X 'q + X t ) V s 0 sin s 0
X t s0
1 '
C2 =
( Xq+ Xt )
X
1
'
C3 = ( X d + X t ) V s 0 cos s0 + R t V s 0 sin s 0
X
1
C 4=Rt
X
1
CV 1= ( X 'q + X t ) cos s 0 + Rt sin s 0
X
1
'
CV 2= Rt cos s 0+ ( X d + X t ) sin s 0
X
C1 =

[
[

s0

Sm0

'
q0

Efd 0

V s0

and

denote the values at the initial operating condition. The

linearized versions of the equations (1) (2) (6) (10) and
(30) are as follows

can be obtained in real time by load ow information at

the transformer and at the generator terminals. So for
any PSS design based on this model, the parameters
can be easily ,

'

T e =K 1 s+ K 2 E q + K v 1 V s
(32)

s=w B S m s
(33)

T mT e D Sm
]
(34)
1
Sm =

2H
K3
E'q =
[ E fdK 4 sK v2 V s ] (
1+ s K 3 T 'd 0

Fig. 5: Linearized model of a single machine in a connected

network

35)

V t =K 5 s + K 6 E'q + K v3 V s
(36)

Modied to accommodate major structural changes in

the system from time to time by local measurements.

'
q

K5 s+ K6 E
KA
E fd =
[ V ref ( + K v 3 V s )]
1+ s T A

(37)

K 1 to K 6 are same as the

The constants
original Heffron-phillips constants (Equations (4)(5) (7)
(8) (11)(12)). However, they are no longer referenced to
and Eb and independent of the equivalent
reactance

Vt

Xe

. They are functions of

Vs

s
Vs

and machine currents. In this model, as

is not a constant, during linearization, three additional
K v 1 to K v 3 are introduced at the
constants
torque, eld voltage and terminal voltage junction
points, that is the major difference between this model
and original Heffron-phillips model.
K v 1 to K v 3 can be written as
'
Eq 0 sin s 0 ( X q X d ) I q 0 cos s 0
K v 1=

( Xt+ Xq)
( X 'd + X t )

(38)

K v 2=

H Control Theory power

system stabilizer design

A method for designing robust power system stabilizer

for a Single Machine Innite Bus (SMIB) power system
is described in this paper. A robust control approach
H control theory is implemented to
based on
provide desired damping to the lightly damped or
unstable mechanical modes. The implemented method
provides adequate damping for the system dynamics
over a range of operating condition. Performance of the
designed controller is simulated over wide range of
operating conditions and compared with conventional
PSS.
Here P(s) is the plant and C is the controller.The signal
w contains all external inputs, including disturbances,
sensor noise, and commands; the output z is an error
signal; y is the measured variables; and u is the control
input. The resulting closed-loop transfer function from
w to z is denoted by T ZW .

( X d X 'd ) cos s 0

( X 'd + X t )

(39)
'

K v 3=

X q V d 0 sin s 0 X d V q 0 cos s0
+
( X q+ X t ) V t 0
( X t + X 'd ) V t 0

(40)
The modied Heffron Phillip's model is shown in the
g 5. If the deviations in the transformer voltage are
neglected then this model exactly represents a system
X e equal to the
with the external reactance
transformer reactance

Xt

. The modied constants

Fig. 6: Basic Block Diagram of

Optimal Control

F =B T2 X

The state-space representation of the above system is

given by

(54)

Z= I

x = Ax+ B1 w+ B2 u
C1 x+ D 12 w

y=

C2 x+ D21 w

(42)
(43)

There exist a compensator C(s) such that

T zw <

X =Ric ( H ) 0
(45)

Y =Ric ( J ) 0
(46)
2

(47)

H =

and

given by
T

B1 B1

T
1

(48)

J =

CT1 C 1

B1 B

T
1

]
]

B 2 BT2

AT

C C1

Robust Power System Stabilizer

The combination of H control technique and the

modied Heffron-philip's model describe to design a
stabilizer. Designed stabilizer is tested with widely
used standard IEEE systems one is 3 generator 9 bus
system. The performance of the designed stabilizer is
compared with conventional stabilizer.
Fig 7 shows single line diagram of 3 generator
9 bus test system. Generator data, network data and
load ow data for the same sytem is given in the
appendix B.

Compensator C(s) exist if and only if

Here

Theory explained in power system stabilizer design has

been used with modied heffron-phillip's model.
Designed stabilizer is tested with widely used standard
IEEE systems one is 3 generator 9 bus system. The
performance of the designed stabilizer is compared
with conventional stabilizer.

(44)

( X Y )<

(55)

(41)

z=

Y X
2

C2 C 2

(49)

Where
is spectral radius of the matrix,
expression for the controller is
C(s) =

A Z L
F
0

By using H control theory PSS has been designed

for all three generator For given test system and their
performance has been compared with conventional
PSS. For performance analysis simulation has been
carried out for nominal operating condition with
creating some disturbances.

(50)
Where

A =A +

B1 B 1 X

(52)
T

L=Y C 2
(53)

+B 2 F +

Fig. 7: 3 Generator 9 Bus System

Z L C 2

where i is the Generator Number

Table 2 gives the undamped or lightly damped rotor
modes of 3 generator 9 bus test system for nominal
operating condition.
Table 1: 3 Generator 9 Bus System : Generator Transfer
Function
Gen.
No.(i)

1
2
3

Fig. 8: Sm1COI for 0.1 pu Change in Tm of Generator 1

n1

d0

d1

d2

d3

-384

11990

3713

309.8

20.94

-246.9

13490

2579

286.3

20.88

-8.245

12510

1126

279.4

20.19

Table 2: 3 Generator 9 Bus System : Undamped or

Lightly damped Rotor Modes
Gen. No

Undamped/ Lightly Damped rotor mode

1
2
3
Fig. 8: Sm1COI for 3

0.007

1.485
1.461

j7.47

j11.14
j13.91

Fault on the Bus no. 7 for the duration

of 100 msec.

Appendix A

: : Rotor angle
s : Rotor angle

with respect to the secondary

voltage of transformer.
S m : Slip speed.

D : Damping coefficient.
E'q , E'd : Transient induced voltages due to eld
Fig. 9: Sm2COI for 0.1 pu Change in Tm of Generator1

ux-linkages.
Efd : Field voltage.

E'dc : Induced voltage of dummy coil used to

account for transient saliency.
H : Inertia Constant of machine.
i d : d-axis component of stator current.

i q : q-axis component of stator current.

K e , K A : Exciter gain.

M : angular momentum.
Pt : Real power injected at the machine terminals.
Q t : Reactive power injected at the machine
Fig.10: Sm2COI for 3

Fault on the Bus no. 7 for the

duration of 100 msec.

The linearized fourth order transfer functions of all

three generators under nominal operating condition
using modied Heffron-phillips's model described in
previous section can be written as

P ( s )=

n1 i s
4

s + d3 i s +d 2 i s +d 1 i s+ d 0 i

(56)

terminals.

Ra : Armature resistance.
T e , T A : Exciter time constant.
T m : Mechanical torque.
T e : Electrical torque.
T 'do , T 'do' : d-axis time constants.
'
''
T qo , T qo : q-axis time constants.
p : power factor at the transformer bus.
T 'c : Time constant of dummy coil used to account
for transient saliency.

V s s : Voltage measured at the secondary of the

transformer.

Table 2: 3 Generator 9 Bus System - Network Data

V t : Voltage measured at the generator

Type

Fro
m

To

B/2

Tap

1
1

2
3

7
9

0
0

0.062
0.058

1
1

0.01

0.085

Appendix B

This Single Machine Innite Bus (SMIB) system data is

taken from [2].

0
0
0.08
8
0.07
9
0.15
3
0.17
9
0.07
4
0.10
4

No of
Parall
el
Elem
ents
0
0

terminal.

V ref : Reference voltage.

V pss : PSS Input.
V d , V q : d and q-axis components of terminal

voltage.

: angular speed.
B : Base Speed
X t , X L : Transformer and transmission line

reactances.
'
d
'
q

Xd , X , X
Xq , X , X

''
d
''
q

: d-axis reactances.
: q-axis reactances.

0.01
7
0.03
2
0.03
9
0.00
8
0.01
1

0.092
0.161
0.17
0.072
0.100
8

Generator Data:

x d=1.6
'
do

T =6
D

x 'd=0.32

EB =1.0 p . u

Efdmax =6

K pss =16

; H = 5;

f B=60 HZ

0;

x q=1.55

x t=0.1 p .u

Static Excitation System Data:

K e =200
T e=0.05 s ;
;

Efdmin =6

PSS Data:

T 1 =0.076
T W =2 ;

T 1 =0.028

0.1

PSS output limits =

Table 1: 3 Generator 9 Bus System - Machine Data

xd xq
0.
14
6
0.
89
5
1.
31
2

'

xd

'

'

xq

Bus No.
1
2
3
4
5
6
7
8
9

Single line diagram for IEEE 3 generator 9 bus test

system is shown in the Fig 7. Base MVA for the system
is 100 MVA and system frequency is 60Hz

Bus
no

Table 3: 3 Generator 9 Bus System - Load ow Data

'

T do T qo

Ra

0.0
96

0.0
60

0.0
96

8.
96

0.3
1

23.
64

0.0
12

0.8
64

0.1
19

0.1
96

0.5
35

6.4

0.0
06

1.2
57

0.1
81

0.2
5

5.
89

0.6

3.0
1

0.0
04

P
0.71641
1.63
0.85
0
-1.25
-0.9
0
-1
0

Q
0.27046
0.066536
-0.1086
0
-0.5
-0.3
0
-0.35
0

V
1.04
1.025
1.025
1.0258
0.99563
1.0127
1.0258
1.0159
1.0324

0
9.28
4.6648
-2.2168
-3.9888
-3.6874
3.7197
0.72754
1.9667

CONCLUSION

Based robust power system stabilizer design

for interconnected power system has been presented.
The use of modied Heffron-Phillip's model [7] permits
the use of H control technique for PSS design of
multi-machine system. PSS at each machine is
synthesized using information available at the local
buses only. The performance of the proposed stabilizer
is reasonably good at all conditions tested.

ACKNOWLEDGMENT
Extending our grateful thanks to the authorities of
Acharya Nagarjuna University for their support and
encouragement to write this paper.

REFERENCES

H00 Control Problems, IEEE Trans. Automatic Control, vol.

[1]

E. V. Larsen and D. Swann, Applying power system

34, No.8, pp. 831-847, June 1989.

stabilizers, parts I, II and III,IEEE Trans. Power App. Syst.,

[2]

vol. PAS-100, No.6, pp. 3017-3046, June 1981.

F. P. Demello and C. Concordia, Concepts of synchronous
machine stability as affected by excitation control, IEEE

[3]

Trans. Power App. Syst., vol. PAS-88, No.4, pp.316-329, 1969.

K.R.Padiyar, Power System Dynamics Stability and

[4]

Control. John Wiley; Interline Publishing, 1996.

P. S. Kundur, Power System Stability and Control. New

[5]

York: McGraw-Hill, Inc., 1994.

P.M.Anderson and A.A.Foud, Power System Control and

[6]

Stability. The Iowa State University Press, 1977.

Gurunath Gurrala, Power System Stabilizing Controllers Multi-machine System, Ph.D Thesis, Indian Institute of

[7]

Science Bangalore, January 2010.

G. Gurunath and I. Sen, A modied Heron-phil1ips
model for the design of power system
stabilizers,
[8]

in

Powercon-2008,

New

Delhi, India, October 12 - 15, 2008.

G. Gurunath and I. Sen, Power system
stabilizers design for interconnected
power systems, IEEE Trans.Power Sys.,

[9]

vol. 25, no.2, pp.10421051,May 2010.

P. Srikant Rao and I. Sen, A QFT Based Robust SVC
Controller for Improving the Dynamic Stability of Power
System ,4 International Conference on Advances in Power
System Control, Operation and Management, APSCOM-97,

Hong Kong, November 1997.

[10] P. Srikant Rao and Indraneel Sen,Robust Tuning of Power
System Stabilizers Using QFT, IEEE Transactions on
Control Systems Technology, Vol. 7, No. 4, pp. 478-486,July
1999.
[11] P. Srikant Rao and Indraneel Sen, Robust pole placement
stabilizer design using linear matrix inequalities , IEEE
Transactions on Power Systems, Vol.15, No. 1, pp. 3003-3008,
Feb. 2000
[12] John C. Doyle, Keith Glover, Promod P. Khargonkar and
Bruce A. Francis, State-Space Solutions to Standard H2 and

[13] V.L.Khartinov, Asymptotic stability of an equilibrium

position of a family of systems of linear differential
equations,Vol. 14, pp. 1483-1485, 1979
[14] Saikat Bhattacharya, L. H. Keel and S. P. Bhattacharyya,
Robust Stabilizer Synthesis for Interval Plant Using H
innity Method , Proceedings of 2 Conference on
Decision and Control, San Antonlo,Texas, December 1993.
[15] A. Hariri and O.P. Malik, A Fuzzy Logic Based Power
System Stabilizer With Learning Abi1ity,IEEE Trans. on
Energy Conuersion,vol. 11, No.4,pp.721727, December
1996.
[16] S. Chen and O.P. Malik, Power System Stabilizer Design
Using it Synthesis,IEEE Trans. on Energy Conversion,vol.
10, No.1,pp.175181, March 1995.

BIBILOGRAPHY
Praveen nayak Bhukya obtained his B.Tech degree in
Electrical and Electronic Engineering from R V R & J C
College of Engineering in 2009 . and M.E degree in
Electrical Engineering from Indian institute of sciences,
Bangalore, in 2011. He is M.Tech student at Acharya
Nagarjuna University College of Engineering and
technology, Guntur, India. His research interests
include power system stability, exible ac transmission
systems, and facts controllers power system stabilizers.