Power System Damping Enhancement via Coordinated Design of

PSS & TCSC in Multimachine Power System

Saleh M. Bamasak * and M. A. Abido **

*Saudi Electricity Company, SEC-ERB Dammam 31422, [email protected], **King Fahd Univ.
of Petroleum & Minerals, KFUPM Dhahran 31261, [email protected] , Saudi Arabia

Abstract — the main objective of this paper is to control strategy for the TCSC where a performance index
investigate the enhancement of power system stability via of time was minimized. A fuzzy logic controller for a
coordinated design of Thyristor Controlled Series TCSC was proposed in [12]. Heuristic optimization
Compensation (TCSC) and Power System Stabilizers (PSSs) techniques have been implemented to search for the
in multimachine power system. The design problem of the
proposed controllers is formulated as an optimization
optimum TCSC based stabilizer parameters for the
problem. Using the developed linearized power system purpose of enhancing SMIB system stability [13]. In
model, the particle swarm optimization (PSO) algorithm is addition, different control scheme for a TCSC were
employed to search for optimal controllers' parameters proposed such as variable structure controller [14-15],
settings that maximize the minimum damping ratio of all bilinear generalized predictive controller [16], and H-
system eigenvalues. The proposed controller is evaluated on based controller [17].
a multimachine power system. The nonlinear simulation A little work has been devoted in the literature to study
results and eigenvalue analysis show the effectiveness of the the coordination control of excitation and FACTS
proposed controller in damping power system oscillations. stabilizers. Hiyama et al [18] presented a coordinated
Index Terms — FACTS, TCSC, PSS, Particle Swarm fuzzy logic-based scheme for PSS and switched series
Optimization (PSO), Power Oscillation Damping (POD),
Multimachine Power System
capacitor modules to enhance overall power system
stability. Robust coordinated design of excitation and
TCSC-based stabilizers using genetic algorithm in SMIB
I. INTRODUCTION also has been presented. Pourbeik and Gibbard [19]
presented a two-stage method for the simultaneous
Power systems are experiencing low frequency
coordination of PSSs and FACTS-based lead-lag
oscillations due to disturbances. These oscillations may
controllers in multimachine power systems by using the
sustain and grow to cause system separation if no
concept of induced damping and synchronizing torque
adequate damping is available [1]. In order to damp
coefficients.
power system oscillation and increase system oscillation
In this paper, a comprehensive assessment of the effects
stability, the installation of power system stabilizer (PSS)
of the coordinated design of PSS and TCSC-based
is both economical and effective [2-5].
stabilizer on power system stability enhancement has
To date, most major electric power system plants in many
been carried out in multimachine power system. The
countries are equipped with PSS. However, PSSs suffer a
controller design problem is transformed into an
drawback of being liable to cause great variations in the
optimization problem where the particle swarm
voltage profile and they may even result in leading power
optimization (PSO) is employed to search for the optimal
factor operation and losing system stability under severe
settings of stabilizer parameters. The location of PSS
disturbances.
selected based on Participation Factor (PF) analysis while
Recently appeared FACTS-based stabilizers such as
Modal Analysis technique is employed to find the TCSC
Static Var Compensator (SVC), Thyristor controlled
location. The eigenvalue analysis and nonlinear
Series Compensation (TCSC), and Thyristor Controlled
simulation results are carried out to demonstrate the
Phase Shifter (TCPS) offer an alternative way in damping
effectiveness of the proposed stabilizers to enhance
power system oscillations. Although, the damping duty of
system stability.
a FACTS controller often is not its primary function, the
capability of FACTS-based stabilizers to increase power
system oscillation damping characteristics has been II. PROBLEM STATEMENT
explored in many aspects [6-19].
Several approaches based on modern control theory have A. Power System Model
been applied to TCSC controller design [9-17]. The
effectiveness of the series compensation devices on A power system can be modeled by a set of nonlinear
power system stability enhancement in SMIB system has differential equations as:
x
been presented in [9]. Chen at al. [10] presented a state X f X ,U (1)
feedback controller for TCSC by using a pole placement Where X is the vector of the state variables and U is the
technique. Cang and Chow [11] developed a time optimal vector of input variables. In this study
X [G , Z , E q' , E fd ]T and U are the PSS and SVC is installed in the feedback loop to generate a stabilizing
output signals. signal uPSS.
In this study, each generator will be presented by the
third-order model comprising of the electromechanical C. Thyristor Controlled Series Compensation (TCSC)
swing equation end the generator internal voltage
equation. Fig. 2 illustrates the block diagram of TCSC with a lead-
x lag compensator. The reactance of XTCSC can be
Gi Z b (Z i  1) (2) expressed as
> @
x 1
x X TCSC K S X ref  U CSC  X TCSC (7)
Zi (T mi  T ei  D i (Z i  1)) / M i (3) TS
x Where:
E qi' ( E fdi  ( x di  x di' ) i di  E qi' ) / T doi
' (4) Xref is the reference reactance of TCSC; Ks and Ts are the
gain and time constant of the device respectively. As
T ei v qi i qi  v di i di (5) shown in Fig. 3 a conventional lead-lag controller is
installed in the feedback loop to generate the
Where, Tm and Te are input and output power of the compensation stabilizer signal UTCSC
generator respectively; M and D are the inertia constant Xmax
and damping coefficient respectively; G and Z are Xref 1
rotor angle and speed respectively; E fd is the field X TCSC
1  sT C

voltage; Tdo' is the open circuit field time constant, xd and Xmin
UTCSC
x d' are the d-axis reactance and d-axis transient
reactance of the generator respectively. Ks
sT w § 1  sT 1
¨¨
· § 1  sT 3 ·
¸¸ ¨¨ ¸¸
'Z
1  sT w © 1  sT 2 ¹ © 1  sT 4 ¹
B. Power System Stabilizers (PSS) Lead-lag Controller

Fig.2: TCSC with Lead-Lag Controller

A PSS can be viewed as an additional block of a
generator excitation control or Automatic Voltage The TSSC has the same block diagram as TCSC except
Regulator (AVR), added to improve the overall power that the minimum limit of the compensation is zero i.e.
system dynamic performance, and especially control Xmin= 0.
electromechanical oscillations. This is a very effective
method of enhancing small-signal stability performance D. Linearized System Model
on a power system network
PSS involves a transfer function consisting of an In the design of electromechanical mode damping
amplification block, a wash out block and two lead-lag controllers, the linearized incremental model around a
blocks. The lead-lag blocks provide the appropriate nominal operation point is employed.
phase-lead characteristic to compensate the phase lag Linearized the system model yield the following state
between the exciter input and the generator electrical equation
torque. The structure of the used PSS is illustrated in Fig. x
1. X AX  HU (8)
V >
Here, the state vector X is ' G , ' Z , ' E q ' , ' E fd @T
max Ecfd
Vref KA And the control vector U is >' X CSC , ' U PSS @ .
1  sT A min

U pss max III. PROPOSED DESIGN APPROACH

sTw §1 sT1PSS ·§1sT3PSS· 'Z
KPSS ¨ ¸¨ ¸
min 1 sTw ¨©1 sT2PSS¸¹¨©1sT4PSS¸¹ In this section the proposed approach is illustrated as
follows; the location of the reactive power compensation
Lead-Lag PSS device TCSC and PSSs are identified in multimachine
Fig. 1: IEEE type-ST1 excitation system with PSS system by using modal analysis method and participation
factor technique respectively. Then, the PSO is proposed
The IEEE Type-ST1 can be described as in this paper to search for optimal parameters setting.
x
E fd K A V ref  v  u PSS  E fd / T A (6) A. Problem Formulation

Where, KA and TA are the gain and time constant of the To increase the system damping to the electromechanical
excitation system respectively; Vref is the reference model, the objective function J defined below is
voltage. As shown in Fig. 1, a conventional lead-lag PSS proposed.
 ­] :] is the min elctromechanical ½ (9) IV. SIMULATION RESULT
J max ® ¾
¯mod e damping ratio ¿
A. Test System
This objective function will identify the minimum value The system considered in this paper is the two-area power
of damping ratio among electromechanical modes of all system. The system one-line diagram is shown in Fig.3.
loading condition considered in the design process. The The details system data including the dynamic generators
design problem can be formulated as the following model and exciter data used along with load flow result
optimization problem format. are given in the Appendix.
The system consists of two identical areas. Each includes
Maximize J two 900 MVA generating units equipped with fast static
Subject to exciters. All four generating units are represented by the
KTCSCmin  KTCSC  KTCSCmax same dynamic model. The power transfer from Area 2 to
Area 1 over a single tie line is considered.
KPSSmin  KPSS  KPSSmax
TTCSC1min TTCSC1  TTCSC1max,
TTCSC3min  TTCSC3  TTCSC3max AREA 1 AREA 2
TPSS1min  TPSS1  TPSS1max , 5 6 7 10 9
8
1 3
TPSSmin  TPSS3  TPSS3max G1 G3

The minimum and maximum value of the controller gain

is set as 0.1 and 100 respectively. The maximum values
of T1 and T3 are set to 1.0s. PSO has been employed to
solve the above optimization problem.

B. Particle Swarm Optimization (PSO) Algorithm 2 Load A Load B 4

17.67MW 9.67MW G4
G2
2.5MVar 2.5MVar
Like evolutionary algorithms, PSO technique conducts
search using a population of particles. Each particle
represents a candidate solution to the problem. In PSO Fig. 3: Two-area 4-machin power system.
System, particles change their positions by flying around
in a multi dimensional search space until a relatively B. System Analysis and Controllers Locations
unchanging position has been encountered, or until
computational limitations are exceeded. In social science From the open loop system eigenvalue and participation
context, a PSO system combines a social-only model and factor analysis shown in Table 1, the system exhibits
a cognition-only model [20]. The social-only component three electromechanical modes:
suggests that individuals ignore their own experience and x An inter-area mode, with a frequency of 0.5098 Hz,
adjust their behavior according to the successful beliefs in which the generating units in one area oscillate
of individuals in the neighborhood. On the other hand, against those in the other area.
the cognition-only component treats individuals as x Local mode, in area 1, with a frequency of 1.1125
isolated beings. The advantages of PSO over other Hz. In this mode the machines in Area 1 oscillate
traditional optimization techniques can be summarized as against each other.
follows: - x Local mode, in area 2, with a frequency of 1.0941
x PSO is a population-based search algorithm i.e., PSO Hz. In this mode the machines in Area 2 oscillate
has implicit parallelism. This property ensures PSO to against each other.
be less susceptible to getting trapped on local minima. The frequencies, damping ratios, and participation factors
x PSO uses objective function information to guide the (PF) for these three electromechanical modes are given in
search in the problem space. Therefore, PSO can easily the table.
deal with non-differentiable objective functions. The table shows that the two generating units in each area
x PSO uses probabilistic transition rules, not have close participation factor in the inter-area mode. The
deterministic rules. Hence, P80 is a kind of stochastic same is also true for the two local modes. This is to be
optimization algorithm that can search a complicated expected, since all units are identical, and units in each
and uncertain area. This makes PSO more flexible and area are electrically close. The table also shows that the
robust than conventional methods. units in Area 1 (the receiving end) have higher
x Unlike GA and other heuristic algorithms, PSO has the participation factor than the units in Area 2 (sending end)
flexibility to control the balance between the global and to the inter-area mode. It can also be seen that, the inter-
local exploration of the search space. area mode has negative damping ratio at this operating
condition.
Table 1: System Eigenvalue and Participation Factor Analysis

Damping Machines Participation Factor

Eigenvalues Freq. Mode
Ratio
G1 G2 G3 G4
-0.660 ±6.9904i 1.1125 Local 0.094 0.7544 1 0.0015 0.0088
-0.7375 ±6.8742i 1.0941 Local 0.1067 0.0133 0.0016 0.8438 1
Inter-
0.0279 ± 3.2030i 0.5098 -0.0087 1 0.7869 0.3891 0.2432
Area

The first electromechanical mode has a very low damping

ratio equal to (0.094) in which Generator no. 1 & 2 have
the significant participation factors of that mode. Table 2: Optimal parameter settings of PSS and TCSC,
Therefore, PSSs are located at machine number 1 and 2 in coordinated design for 4-machine system
addition to machine 4 since it has the significant PF of
the inter-area mode. Coordinated Design
Parameters
The TCSC is to be installed at the tie-line. This location PSS1 PSS2 PSS4 TCSC
is satisfied the primary function of TCSC as will as the K 100 100 49.2614 1.064
practical experience.
T1 0.0783 0.0702 0.1354 5.0
C. Controller Design T2 0.01 0.01 0.01 0.021
T3 ------ ------ ------ 0.01
All stabilizers PSS's & TCSC are simultaneously tuned T4 ------ ------ ------ 5.0
by PSO searching for the optimum controllers parameter
settings that maximize the minimum damping ratio of all
D. Eigenvalue Analysis
the system complex eigenvalues.
The convergence rate of the objective function when
The system eigenvalues with the proposed PSS's and
PSS's and TCSC-based controllers are designed
TCSC-based stabilizers when applied individually and by
individually and in a coordinated manner is shown in Fig.
means of coordinated design is given in Table 3. The bold
4. It is clear that the coordinated design of PSS's and
rows of this table represent the EM modes eigenvalues
TCSC-based stabilizer improves greatly the system
and their damping ratios and frequency. It is evident that,
damping compared to their individual application. The
using the proposed coordinated stabilizers design, the
final settings of the optimized parameters for the
damping ratio of the EM mode eigenvalue is greatly
proposed stabilizers are given in Table 2.
enhanced. Hence, it can be concluded that this improves
the system stability.

Table 3: System eigenvalues with the proposed stabilizer

PSSs TCSC TCSC & PSSs

-3.9219±5.7975i -0.749±6.8431i -4.423±6.07504i
0.5603*, 0.922** 0.108*, 1.09** 0.59*, 0.97**
-3.4337±5.4358i -0.509±4.3648i -6.16589±5.0236i
0.534*, 0.8651** 0.115*, 0.69** 0.783*, 0.8**
-1.7682±1.6360i -2.390±3.2546i
-0.597±5.0238i
0.734*, 0.2604** 0.59*, 0.526**
-2.6113±3.0188i -1.614±4.9672i -1.9658±2.7841i
-7.7989±11.9581i -89.067,-89.41 -6.4785±9.132i
-12.372±17.9649i -79.4 -13.00±15.812i
-17.3096±0.1076i -76.692,-23.969 -22.269±4.7965i
-21.5036±1.8633i -21.044,-16.712 -100,-100,-0.2,
-92.0013,-89.364 -13.890,-13.650 -100, -128.36,
-81.0185,-76.811 -7.5208,-6.4332 -82.485,-100
-11.898, -7.4442, -5.4079, -4.908 -20.009,-17.72,
Fig. 4: The convergence rate of the objective function when
-6.2647, -4.9271, -0.2000 -15.80, -12.841,
PSS's and TCSC-based controllers are designed individually
and in a coordinated manner -0.2024, -0.2122 -9.4599,-6.023,
-5.3289,-2.619
* damping ration, ** frequency (Hz)
E. Nonlinear Time Domain Simulation effectiveness of the proposed stabilizer and its ability to
provide good damping of low frequency oscillation and
Figs. 5-6 show the speed deviations and the rotors angle improve greatly the system voltage profile.
responses, for a 6-cycle 3 fault at bus 7 while using the
proposed PSSs and TCSC coordinated design.
The simulation results obtained clearly indicate that the VI. ACKNOWLEDGMENT
proposed coordinated PSS-TCSC design enhances the The first author would like to acknowledge the support of
system stability. These results confirm the conclusion Saudi Electricity Company-ERB.
drawn from eigenvalue analysis results. The second author would like to acknowledge the support
of King Fahd University of petroleum & Minerals.

VII. REFERENCES
[1] Y. N. Yu, Electric Power System Dynamics,
Academic Press, 1983.
[2] F. deMello and C. Concordia, “Concepts of
Synchronous Machine Stability as Affected by
Excitation Control,” IEEE Trans. PAS, Vol. 88, pp.
316-329, 1969.
[3] G. T. Tse and S. K. Tso, "Refinement of
Conventional PSS Design in Multimachine System
by Modal Analysis," IEEE Trans. PWRS, Vol. 8, No.
2, 1993, pp. 598-605.
[4] Y.Y. Hsu and C.Y. Hsu, “Design of a Proportional-
Integral Power System Stabilizer,” IEEE Trans.
PWRS, Vol. 1, No. 2, pp. 46-53, 1986.
[5] M. A. Abido and Y. L. Magid, “A hybrid neuro-
fuzzy power system stabilizer for multimachine
Fig 5: Speed response for 6-cycle fault with PSS & power systems,” IEEE Trans. on PWRS, Vol. 13, No.
TCSC, coordinated design 4, November 1998, pp. 1323-1230.
[6] N. G. Hingorani, "Flexible AC Transmission," IEEE
Spectrum, April 1993, pp. 40-45.
[7] Y. Wang, R. Mohler, R. Spee, and W. Mittelstadt,
"Variable Structure FACTS Controllers for Power
System Transient Stability," IEEE Trans. PWRS,
Vol. 7, 1992, pp. 307-313.
[8] R. Byerly, and E. Taylor, "Static Reactive
Compensation for Power Transmission System,"
IEEE Trans. PAS-101, 1982, pp. 3998-4005
[9] S. M. Bamasak and M. A Abido, “Effectiveness Of
Series Compensation On Power System
Stability Enhancement.” GCC-Cigre conf., Vol. 1, pp
65-71, Nov 2004.
[10] X. Chen, U. Annakkage, and C. Kumble, "Controlled
Series Compensation for Improving the Stability of
Multimachine Power Systems," IEE Proc., Pt. C,
Vol. 142, 1995, pp. 361-366
[11] J. Chang and J. Chow, "Time Optimal Series
Fig 6: Rotor angle response for 6-cycle fault with PSS & Capacitor Control for Damping Inter-Area Modes in
TCSC, individual and coordinated design Interconnected Power Systems," IEEE Trans. PWRS,
Vol. 12, No. 1, 1997, pp. 215-221.
V. CONCLUSION [12] T. Lie, G. Shrestha, and A. Ghosh, "Design and
Application Of Fuzzy Logic Control Scheme For
In this study, a coordination design of TCSC and PSS Transient Stability Enhancement In Power System",
stabilizers is proposed. The tuning parameters of the Electric Power System Research. 1995, pp, 17-23.
proposed stabilizer were optimized using PSO. The [13] M. A. Abido, “Genetic-based TCSC damping
proposed stabilizer have been applied and tested on a controller design for power system stability
weakly connected multimachine power system under enhancement,” Electric Power Engineering, 1999.
severe disturbance. The eigenvalues analysis and the [14] PowerTech Budapest 99. International Conference
nonlinear time domain simulation results show the on , 29 Aug.-2 Sept. 1999, Page(s): 165
[15] Y. Wang, R. Mohler, R. Spee, and W. Mittelstadt, Table 5: 4-machine System line data in per unit.
"Variable Structure FACTS Controllers for Power Line
From To R X B
System Transient Stability," IEEE Trans. PWRS, no.
Vol. 7, 1992, pp. 307-313. 1 1 5 0 0.0167 0
[16] T. Luor and Y. Hsu, "Design of an Output Feedback 2 2 6 0 0.0167 0
3 3 8 0 0.0167 0
Variable Structure Thyristor Controlled Series
4 4 9 0 0.0167 0
Compensator for Improving Power System 5 5 6 0.0025 0.025 0.021875
Stability," EPSR, 47, 1998, pp. 71-77. 6 8 9 0.0025 0.025 0.021875
[17] V. Rajkumar, R. Mohler, "Bilinear generalized 7 6 7 0.001 0.01 0.00875
predictive control using the TCSC," IEEE Trans. 8 9 10 0.001 0.01 0.00875
PWRS 9 (4) (1994) 1987-1993.
[18] Q. Zhao and J. Jiang, "A TCSC Damping Controller Table 6: Machine Data
Using Robust Control Theory," Int. J. of Elec. Power M H Xd Xd' Xq Xq' Td'
& Energy Systems, Vol. 20, No. 1, 1998, pp. 25-33. 1 55.575 0.2 0.033 0.19 0.016 8
[19] T. Hiyama, M. Mishiro, H. Kihara, and T. Ortmeyer, 2 55.575 0.2 0.033 0.19 0.016 8
"Coordinated Fuzzy Logic Contrl for Series 3 58.5 0.2 0.033 0.19 0.016 8
Capacitor Modules and PSS to Enhance Stability of
4 58.5 0.2 0.033 0.19 0.016 8
Power System," IEEE Trans. PWRD, Vol. 10, No. 2,
Tq' KA TA D
1995, pp. 1098-1104.
[20] P. Pourbeik, M. J. Gibbard, “Simultaneous 0.4 200 0.01 0
coordination of power system stabilizers and FACTS 0.4 200 0.01 0
device stabilizers in a multimachine power system 0.4 200 0.01 0
for enhancing dynamic performance,” IEEE 0.4 200 0.01 0
Transactions on PWRS, Vol.: 13 Issue: 2 , May
1998, pp: 473 -479 Table 7: Load flow result of the 4-machine system
[21] J. Kennedy, “The Particles Swarm: Social Bus Angle Load Generation
Adaptation of Knowledge,” Proceedings of the 1997 Voltage
no. (degree) P Q P Q
IEEE international Conference on Evolutionary
1 1.03 0 0 0 7.2532 2.8008
Computation ICE’97, Indianpolis, Indiana, USA 2 1.01 -10.65 0 0 7 4.4762
1997, pp. 303-308 3 1.03 27.292 0 0 7 1.7721
4 1.01 17.548 0 0 7 2.155
Appendix A 5 0.99159 -6.8112 0 0 0 0
6 0.94312 -17.7 0 0 0 0
4-Machine Power System Data: 7 0.89954 -27.012 17.67 2.5 0 0
8 1.0077 20.825 0 0 0 0
Table 4: 4-machine System bus data in per unit value. 9 0.98122 10.774 0 0 0 0
Load Generation 10 0.96662 2.4251 9.67 1 0 0
Bus
Type Voltage Angle
no. P Q P Q
1 1 1.03 0 0 0 0 0
2 2 1.01 0 0 0 7 0
3 2 1.03 0 0 0 7 0
4 2 1.01 0 0 0 7 0
5 3 1 0 0 0 0 0
6 3 1 0 0 0 0 0
7 3 1 0 17.67 2.5 0 0
8 3 1 0 0 0 0 0
9 3 1 0 0 0 0 0
10 3 1 0 9.67 1 0 0