A Modified Heffron-Phillip’s Model for The Design

of Power System Stabilizers

Gurunath Gurrala, and Indraneel Sen

Abstract—In this paper a modified Heffron-Phillip’s (K- of the power system. Even in the case of single machine
constant) model is derived for the design of power system infinite bus models, estimates of equivalent line impedance
stabilizers. A knowledge of external system parameters, such and the voltage of the remote bus are required. The PSS
as equivalent infinite bus voltage and external impedances or
their equivalent estimated values is required for designing a design also requires information of the rotor angle δ measured
conventional power system stabilizer. In the proposed method, with respect to the remote bus. These parameters cannot be
information available at the secondary bus of the step-up trans- measured directly and need to be estimated based on reduced
former is used to set up a modified Heffron-Phillip’s (ModHP) order models of the rest of the system connected to the
model. The PSS design based on this model utilizes signals generator. If the available information for the rest of the system
available within the generating station. The efficacy of the pro-
posed design technique and the performance of the stabilizer has is inaccurate, the conventionally designed PSS may result in
been evaluated over a range of operating and system conditions. poor system performance.
The simulation results have shown that the performance of the The method proposed for the PSS design in this paper is
proposed stabilizer is comparable to that could be obtained by also based on the classical design technique. However, as
conventional design but without the need for the estimation and opposed to a conventional stabilizer, the proposed PSS judges
computation of external system parameters. The proposed design
is thus well suited for practical applications to power system system disturbances such as changes in system configuration
stabilization, including possibly the multi-machine applications or variation in loads etc, based on the deviations in power flow,
where accurate system information is not readily available. voltage and voltage angle at the secondary bus of the step-up
Index Terms—Heffron-Phillip’s model, Power System Stabiliz- transformer. The PSS tries to control the rotor angle measured
ers(PSS), with respect to the local bus rather than the angle δ measured
with respect to the remote bus to damp the oscillations.
All PSS design parameters are thus calculated from local
I. I NTRODUCTION measurements and there is no need to estimate or compute the
NE of the major problems in power system operation is
O related to the small-signal oscillatory instability caused
by insufficient natural damping in the system. The most cost-
values of equivalent external impedances, bus voltage and rotor
angles at the remote bus. The performance of the proposed
stabilizer is comparable to that of a conventional stabilizer
effective way of countering this instability is to use auxiliary that has been designed based on accurate system information.
controllers called power system stabilizers (PSS), to produce This information is not always available in practical systems.
additional damping in the system [1], [2]. Effective PSS design The paper consists of three parts: the first part describes the
for large electric power systems is extremely laborious because modelling of the power system, the second part describes
of their highly nonlinear nature and constantly changing gen- the modified Heffron-Phillip’s model and the proposed PSS
eration, transmission, and loading conditions. Over the years design procedure and the third part describes the dynamic
a variety of design procedures and algorithms [3] have been performance of the PSS over a range of operating and system
proposed for the design of power system stabilizers using both conditions.
linearized and nonlinear models of power system. However,
because of complex structures and real time computational re-
II. M ODELING OF P OWER S YSTEM
quirements, most of these stabilizers have found little practical
application. For small-signal stability analysis, dynamic modeling is
The concept of classical PSS and their tuning procedures are required for the major components of the power system. It
well explored in [1], [2]. The conventional fixed gain stabiliz- includes the synchronous generator, excitation system, auto-
ers perform reasonably well if they have been tuned properly matic voltage regulator (AVR) etc. Different types of models
[4]. Though these stabilizers have simple robust structures, have been reported in the literature depending upon their
tuning them not only requires considerable expertise but also specific application. A Single Machine Infinite Bus (SMIB)
a knowledge of system parameters external to the generating power system model as shown in fig.1 is used to obtain
station. These parameters may vary during normal operation the linearized dynamic model [5] (Heffron Phillip’s or K-
constant model). Here, a single generator represents a single
machine equivalent of a power plant (consisting of several
Gurunath.G and Indraneel Sen are with the Department of Electrical
Engineering, Indian Institute of Science, Bangalore, India, 560012. Ph.91- generators). The generator is connected to a single or double
080-22932363,22932931, [email protected], [email protected] circuit line through a transformer. The line is connected to
978-1-4244-1762-9/08/$25.000 2008IEEE
 necessary steps to arrive at the modified HP model are given.
From model 1.0 the following equations can be obtained
Eq + Xd id − Ra iq = Vq
(7)
−Xq iq − Ra id = Vd
The subscripts q and d refers to the q and d-axis respectively
in Park’s reference frame. The machine network interface is
achieved by converting machine quantities in Park’s frame to
Fig. 1. A Single Machine Power System Model. synchronously rotating Kron’s reference frame. The machine
terminal voltage in terms of the transformer secondary is given
by
the rest of the power system which may be an infinite bus VQ + jVD = (Vq + jVd )ejδ
or another machine. The infinite bus, by definition, represents
= (iq + jid )(Rt + jXt )ejδ + Vs ∠θs
a bus with fixed voltage source. The magnitude, frequency
and phase of the voltage are unaltered by changes in load
(output of the generator). This is a simplified representation ∴ (Vq + jVd ) = (iq + jid )(Rt + jXt ) + Vs ∠θs e−jδ
of a remote generator connected to a load center through
a transmission line. IEEE Model 1.0 is used to model the Replacing δ by δs + θs in the above equations gives
synchronous generator [6] with a high gain, low time constant
static exciter. The dynamic equations governing the system are
as follows. (Vq + jVd ) = (iq + jid )(Rt + jXt ) + Vs ∠ − δs
δ̇ = wB Sm (1) Equating the real and imaginary parts of the above equation
gives
1 Vq = Rt iq − Xt id + Vs cos δs
Ṡm = {Tmech − Telec − DSm } (2) (8)
2H Vd = Rt id + Xt iq − Vs sin δs
substituting (8) in (7) and rearranging gives
1  
     
Ėq = − Eq + (Xd − Xd )id + Ef d (3)

Tdo Xd + Xt −Rt id Vs cos δs − Eq
= (9)
−Rt Xq + Xt iq −Vs sin δs
1
Ėf d = {− Ef d + Ke (Vref + Vpss − Vt )} (4) The system mechanical equations, electrical equations and
Te
eqn.(9) are linearized as in [6] to obtain the following modified
Telec = Eq iq + (Xd − Xq )id iq K-constants.
Vs0 Eq0 cos δs0 Xq −Xd
K1 = Xq +Xt + Xt +Xd Vs0 sin δs0
The variables have standard meaning and are listed in the
Appendix. The above equations are based on rotor angle δ Xq +Xt
K2 = Xt +Xd iq0 ;
measured with respect to the remote bus Eb . To get the
dynamic equations with respect to the secondary bus voltage
Xt +Xd
Vs ∠θs of the step up transformer, all the expressions involving K3 = Xd +Xt ;
the rotor angle δ have to be expressed in terms of δs , where
δs = δ − θs . The expressions for δs and Eq are as under Xd −Xd
K4 = Xt +Xd Vs0 sin δs0 ;
Ps (Xt +Xq )−Qs Ra
δs = arctan Ps Ra +Qs (Xt +Xq )+Vs2 Xq Vd0 Vs0 cos δs0 Xd Vq0 Vs0 sin δs0
K5 = − (Xq +Xt )Vt0 −
(5) (Xt +Xd )Vt0
if δs < 0 then δs = π − |δs |
Xt Vq0
K6 = Xt +Xd Vt0 ;
  2
(Xt +Xd )
Eq
Xq
= Xt Vt2 − (Xt +Xq ) Vs sin δs Kv1 =
Eq0 sin δs0
−
(Xq −Xd )Iq0 cos δs0
(6) (Xt +Xq ) (Xd +Xt )
X
− Xdt Vs cos δs
(Xd −Xd ) cos δs0
Kv2 = −
III. M ODIFIED H EFFRON -P HILLIPS M ODEL AND PSS (Xd +Xt )
DESIGN
X V sin δ Xd Vq0 cos δs0
Kv3 = − (Xq q +X
d0 s0
t )Vt0
+
The standard Heffron Phillips model can be obtained by lin- (Xt +Xd )Vt0
earizing the system equations around an operating condition. 
The development of the model is detailed in [6]. Here only the where Eq0 = Eq0 − (Xq − Xd ) id0
 K1
PSS Ȧ(s)
ǻTe1
ǻTm 1 ǻȦ ȦB ǻįs

2Hs s
ǻTe2 ǻșs
D
K5
K4
KV1 K2
K3 KE
ǻE'q 1+sTIdo 1+sTE ǻEref
ǻEfd Fig. 3. Structure of PSS.
ǻEt
K6

GEP (s)
crossover of 900 point occurs beyond 22 rad/sec (3.5Hz) [2],
KV2 [6]. The gain of the PSS is selected from the root-locus plot
ǻVs to give maximum damping to the concerned mode of the
KV3 generator and any other (e.g. exciter mode) modes. The form
of the compensator is assumed as given below
m
Fig. 2. Modified Heffron-Phillips model, the rotor angle is Δδs (1 + sT1 )
H(s) = Kpss
(1 + sT2 )
where m is the number of lead-lag stages. The constants T1
The modified Heffron-Phillip’s model comprises six con- and T2 can be obtained from the following equations.
stants K1 to K6 whose definitions remain unchanged. How-
β
ever, they are no longer referenced to δ and Eb . It can be T2 1−sin( m )
α= T1 = β
1+sin( m
observed that the modified K-constants are also no longer the )
functions of the equivalent reactance Xe . They are functions
of Vs , δs , Vt and machine currents. Therefore the modified K- β = Required phase compensation
constants can be now computed based on local measurements
1√
only. In this model, as Vs is not a constant, during linearization, T1 = Ωi α
three additional constants Kv1 to Kv3 are introduced at the
torque, field voltage and terminal voltage junction points as T2 = αT1
shown in fig.2. The action of the PSS is effective through the
where Ωi is the frequency of the mode of interest.
transfer function block GEP (s) as shown in fig.2 between the
electric torque and the reference voltage input with variation
in the machine speed assumed to be zero. IV. S IMULATION R ESULTS AND OBSERVATIONS
The expression for the transfer function GEP (s) is given The performance of the stabilizers designed by using mod-
by ified K-constants is evaluated on a SMIB test system over a
K2 K3 EXC (s) range of operating conditions as shown in table I. The system
GEP (s) =  K ) + K K EXC (s) (10) data is given in the Appendix. Conventional PSS is designed
(1 + sTdo 3 3 6
following the tuning guidelines [2] for Xe = 0.4p.u. The PSS
where EXC(s) is the transfer function of the excitation data for both the conventional design and the proposed method
system. It can be of any exciter, but in this paper a high are also given in the Appendix. The transformer reactance Xt
gain, low time constant static exciter is assumed. As is 0.1p.u. The total impedance between the generator bus and
system operating conditions change, the gain and phase the infinite bus, denoted by Xe varies with system conditions.
characteristics of the transfer function GEP (s) change. Fig.4 shows the phase plots of GEP (s) with modified HP
Ideally, the PSS transfer function should be reciprocal of and conventional HP models. For the test system the center
GEP (s) for providing a prescribed amount of damping with frequency is chosen as 3.5 Hz. Fig.5 shows the phase plots of
speed input. This would be purely a lead function that is conventional (CPSS) and proposed PSS and the compensated
not physically realizable. A practical approach is to have a GEP (s) of the plant in both cases. It is evident that the
lead-lag circuit that provides adequate compensation over proposed PSS achieves exact compensation for the desired
the desired range of frequencies. Using the modified K- range of frequencies (0.1Hz to 2.5Hz).
constants, stabilizers are designed using the tuning guidelines Fig.6 shows the root locus plot of the plant by varying
given by [2]. The stabilizer considered is a simple lead-lag PSS gain with the proposed and conventional PSS. The gain
compensator as shown in fig.3, with a washout filter. The Kpss is chosen as 13 for proposed PSS and 16 for CPSS.
time constants are selected such that the compensated phase It can be observed from the figure that the chosen weights
lag of GEP (s) × P SS(s) around local mode frequency provide adequate damping for both the rotor and the exciter
(about 7 rad/s i.e. 1.12 Hz is assumed) lies below 450 and modes. The performance of the proposed PSS was tested at
 TABLE I
R ANGE OF O PERATING CONDITIONS FOR SMIB

Xe Pt Qt power factor
0.4-Nominal 1, 0.8, 0.8 0.2, 0.2, -0.2 lag, lag, lead
0.3-strong all 0.8 0.41,0.23,-0.37 lag, lag, lead
0.8-weak 1, 0.8 0.5, 0.2 lag, lag

Fig. 5. compensated GEP(s) plots with Proposed PSS - and CPSS –

Fig. 4. GEP(s) plots for modHP −, original HP −−

varying operating and system conditions. A few representative

examples have been included in this paper.
Fig.7 shows the system response in terms of variation in
slip speed Sm following a 10% step change at Vref input of
the generator. At this operating condition (S = P + jQ =
1 + j0.2p.u.,Xe = 0.4p.u) the system is unstable without a
PSS. Fig.8 shows the system response for the same system
condition, following a 3φ fault of 4 cycles duration at the
transformer bus. Fault is cleared by tripping one of the parallel
lines. In both the cases, the conventional and the proposed PSS
have damped the system oscillations effectively.
Fig.9 relates to leading power factor operation with S =
0.8 − j0.2p.u. and Xe = 0.4p.u. System behavior is highly Fig. 6. root-locus plot of the plant with proposed PSS and CPSS.
oscillatory in this case for a 10% step change at Tm input of
the generator. The performance of the proposed PSS is much
better than the conventional stabilizer under this condition.
Fig.10 shows system response in terms of Sm under rel-
atively strong system ( Xe = 0.3p.u., S = 0.8 − j0.37 )
and leading power factor conditions. The proposed PSS has
shown comparable performance under lagging power factor
conditions and better performance under leading power factor
conditions when compared to the performance of the CPSS.
Fig.11 depicts very weak system (Xe = 0.8p.u., S = 1 +
j0.5 p.u. ) conditions. Leading power factor operations are
not possible under these conditions. The performance of both
stabilizers are again comparable and the system oscillations
have been effectively damped.

V. C ONCLUSIONS
A modified Heffron Phillip’s model has been derived for the Fig. 7. System response for 10% step change in Vref ,Nominal system.
design of power system stabilizers. The stabilizer is synthe-
sized using information available at the local buses and makes
Fig. 8. System response for a 3φ fault at transformer,Nominal system. Fig. 11. System Response for 10% step change in Vref , Weak System.

to multi-machine systems.

A PPENDIX
Machine Data:
Xd = 1.6; Xq = 1.55; Xd = 0.32; Tdo 
= 6; H = 5; D =
0; fB = 60Hz; EB = 1p.u.; Xt = 0.1; Model 1.0 is
considered for the synchronous machine.
Exciter data:
Ke = 200; T e = 0.05s; Ef dmax = 6p.u.; Ef dmin =
−6p.u.;
CPSS data:
T1 = 0.078; T2 = 0.026; Kpss = 16; Tw =
Fig. 9. System response for 10% step change in Tm ,Nominal system, leading 2; P SS output limits ± 0.05
p.f. ModHP-PSS data:
T1 = 0.0952; T2 = 0.0217; Kpss = 13; Tw =
2; P SS output limits ± 0.05
no assumptions about the rest of the system connected beyond variables definitions:
the secondary bus of the step up transformer. As system δ : Rotor angle.
information is generally not accurately known or measurable δs : Rotor angle with respect to the secondary voltage of
in practice, the proposed method of PSS design is well suited transformer.
for designing effective stabilizers at varied system conditions. Sm : Slip speed.
The performance of the proposed stabilizer is comparable Tmech and Telec : Mechanical and Electrical torques
to that of a conventional stabilizer which has been designed respectively.
assuming that all system parameters are known accurately. As D : Damping coefficient.
the proposed design is based on local measurements alone it Eq : Transient emf due to field flux-linkage.
may be possible to extend the proposed PSS design philosophy id : d-axis component of stator current.
id : q-axis component of stator current.

Tdo : d-axis open circuit time constant.
Xd , Xd : d-axis reactances.
Xq , Xq : q-axis reactances.
Ef d : Field voltage.
Ke , Te : Exciter gain and time constant.
Vt : Voltage measured at the generator terminal.
Vs : Voltage measured at the secondary of the transformer.
Vref : Reference voltage.
Vpss : PSS input.
Xt , XL : Transformer and transmission line reactances.

R EFERENCES
[1] F.P.Demello and C.Concordia, “Concepts of synchronous machine stabil-
Fig. 10. System response for 10% step change in Vref , Strong system, ity as affected by excitation control,” IEE Trans. Power Apparatus and
leading p.f. Systems, vol. PAS-88, No.4, pp. 316–329, 1969.
[2] E. Larsen and D. Swan, “Applying power system stabilizers,parts I,II Indraneel Sen received his Ph.D degree from IISc, Bangalore, in 1981. He is
and III,” IEEE Trans.Power Apparatus and Systems, vol. PAS-100, pp. currently an Associate professor in the Department of Electrical Engineering
3017–3046, June 1981. at the Indian Institute of Science. His research interests include power system
[3] B. Pal and B. Chaudhuri, Robust Control in Power Systems. Springer, stability, adaptive control and energy management systems.
2005.
[4] G. Rogers, Power System Oscillations. Kluwer Academic Publishers,
2000.
[5] W. G. Heffron and R. A. Phillips, “Effect of modern amplidyne voltage
regulators on underexcited operation of large turbine generators.” Amer-
ican Institutions of Electrical Engineers, vol. 71, pp. 692–697, 1952.
[6] K.R.Padiyar, POWER SYSTEM DYNAMICS Stability and Control. John
Wiley; Interline Publishing, 1996.

Gurunath Gurrala received his B.Tech degree in Electrical and Electron-

ics Engineering from S.V.H.College of Engineering, Nagarjuna University,
Andhrapradesh, in 2001. M.Tech degree in Electrical Power Systems from
J.N.T.U. college of Engineering, Anantapur, A.P., in 2003. He is currently
a research student working towards his Ph.D in Indian Institute of Science,
Bangalore. He worked as an Assistant Professor in Anil Neerukonda Institute
of Technology and Sciences, Visakhapatnam, A.P., during 2003-2005. His
research interests include Power system stability, Flexible AC Transmission
Systems, Non Linear and Adaptive Control of Power systems.