Bookchapter Fke40

First Edition 2008
AZHAR KHAIRUDIN 2008
Hak cipta terpelihara. Tiada dibenarkan mengeluar ulang mana-mana bahagian artikel, ilustrasi, dan isi kandungan buku ini dalam apa juga bentuk dan cara apa jua sama ada dengan cara elektronik, fotokopi, mekanik, atau cara lain sebelum mendapat izin bertulis daripada Timbalan Naib Canselor (Penyelidikan dan Inovasi), Universiti Teknologi Malaysia, 81310 Skudai, Johor Darul Tazim, Malaysia. Perundingan tertakluk kepada perkiraan royalti atau honorarium. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical including photocopy, recording, or any information storage and retrieval system, without permission in writing from Universiti Teknologi Malaysia, 81310 Skudai, Johor Darul Tazim, Malaysia.
Perpustakaan Negara Malaysia
Cataloguing-in-Publication Data
Development of power system dynamic for damping oscillation / editor: Azhar Khairudin. ISBN 978-983-52-0678-8 1. Electric power systems. I. Azhar Khairuddin, 1963-. 621.319
Editor: Azhar Khairudin Pereka Kulit: Mohd Nazir Md. Basri & Mohd Asmawidin Bidin Diatur huruf oleh / Typeset by
Fakulti Kejuruteraan Elektrik
Diterbitkan di Malaysia oleh / Published in Malaysia by

PENERBIT UNIVERSITI TEKNOLOGI MALAYSIA
34 38, Jln. Kebudayaan 1, Taman Universiti, 81300 Skudai, Johor Darul Tazim, MALAYSIA. (PENERBIT UTM anggota PERSATUAN PENERBIT BUKU MALAYSIA/ MALAYSIAN BOOK PUBLISHERS ASSOCIATION dengan no. keahlian 9101) Dicetak di Malaysia oleh / Printed in Malaysia by
UNIVISION PRESS SDN. BHD.
Lot. 47 & 48, Jalan SR 1/9, Seksyen 9, Jalan Serdang Raya, Taman Serdang Raya, 43300 Seri Kembangan, Selangor Darul Ehsan, MALAYSIA.
Contents
CONTENTS
Preface vii
1 Chapter1 Comparison of LQR and Lead Lag Controller for Tuning Power System Stabilizer for Damping Oscillations Nuraddeen. Magaji Mohd Wazir Mustafa Chapter 2 Eigenvalues Based Method for Analysis of Power System with SVC Device for Damping Oscillations Nuraddeen. Magaji Mohd Wazir Mustafa Chapter 3 Damping controller Design for SVC Device using Residue Factor Mohd Wazir Mustafa Nuraddeen. Magaji Chapter 4 Optimal Location of FACTS devices for Damping oscillations Mohd Wazir Mustafa Nuraddeen. Magaji Faridah Mohd Taha 21
41
59
vi
Contents
Chapter 5 Eigenvalues Based Method for Analysis of Power System with TCSC Device for Damping Oscillations Mohd Wazir Mustafa Nuraddeen Magaji Zaniah Muda
79
Chapter 6 Application of FACTS Devices for Damping Out Power Systems Oscillations Using Model Techniques Nuraddeen. Magaji Mohd Wazir Mustafa Ahmad Safawi Mokhtar INDEX
102
121
PREFACE
All praises and thanks are due to Allah SWT who has given us opportunity and provided us with all the necessary means to complete this work. The aim of this manuscript is to make a comprehensive presentation of recent research into the application of linear control system in dynamic of power system to the damping of inter-area oscillations using power systems stabilizer and FACTS devices. The subject is introduced with application of pole placement approach in designing of power system stabilizers (PSS). The main purpose of this manuscript is to documents an idea of damping oscillation using PSS and FACTS devices in a large power system, for engineers engaged in the research, design and development of power systems with particular concern for power system stability. In chapter 1 application of Linear Quadratic Controller (LQR) is used to compare with Lead Lag Controller in designing PSS to damp local and inter-area oscillations. Chapter 2 focused .on the analytic tools that are commonly used in small-signal stability analysis base on eigenvalues such as Controllability index, Observability index, Residue factor and Participation factor for two case studies with Static VAr Compensators (SVC). Chapter 3, a Residue factor approach giving eigenvalue is used to design a power oscillation damping (POD) controller for 11 bus system with SVC device. The method addresses the
viii
Preface
robustness of the control schemes. The systems model which is used to test damping controller design is also described. In chapter 4 the problem of placement of FACTS devices for damping oscillation is addressed based on Residue Factor approach. The optimal location is obtained on different operating conditions in terms of (N-1) contingency. The FACTS controllers consider are: Thyristor-controlled series capacitors (TCSC), Static VAr Compensators (SVC) and Unified Power Flow Controller (UPFC).While 11-bus system and New England 39-bus system were considered as case studies. Analyses of Power System with TCSC Device for Damping Oscillations are detailed in chapter 5. In this chapter eigenvalue analysis were carried out with the TCSC placed at strategic location for SIMB and 11 bus system. In Chapter 6 model Controllability index and Observability Index and their measurement are explained in this chapter. Detailed calculations of these indices for TCSC, SVC and UPFC were also explained. While application of this concepts for determine best location were demonstrated in 16 machines,68 bus system which is equivalent of the interconnected New England test system (NETS) and New York power system (NYPS). In all the above designs, power oscillation damping for variable power level and changing network topology is confirmed by eigenvalue analysis and time-domain nonlinear simulations have been made to demonstrate the power of the designs.
Azhar Khairuddin Faculty of Electrical Engineering Universiti Teknologi Malaysia 2008
Comparison of LQR and Lead Lag Controller for Tuning Power System for Damping Oscillations
1
COMPARISON OF LQR AND LEAD LAG CONTROLLER FOR TUNING POWER SYSTEM STABILIZER FOR DAMPING OSCILLATIONS
N. Magaji M.W. Mustafa
1.1 INTRODUCTION Power systems experience low-frequency oscillations due to disturbances. These low frequency oscillations are related to the small signal stability of a power system. The phenomenon of stability of synchronous machine under small perturbations is explored by examining the case of a single machine connected to an infinite bus system (SMIB). The analysis of SMIB [4] gives physical insight into the problem of low frequency oscillations. These low frequency oscillations are classified into local mode, inter area mode and torsional mode of oscillations. The SMIB system is predominant in local mode low frequency oscillations. These oscillations may sustain and grow to cause system separation if no adequate damping is available. In recent years, modern control theories have been applied to power system stabilizer (PSS) design problems. These include optimal control,adaptive control, variable structure control, and intelligent control. Despite the potential of modern control techniques with different structures, power system utilities still prefer the conventional leadlag PSS structure. The reasons behind that might be the ease of online tuning and the lack of assurance of the stability related to some adaptive or variable structure techniques.
Developments of Power System Dynamic for Damping Oscillation
The main objective of this chapter is to evaluate a control technique, to design a damping controller for PSS devices. This work uses LQR and Lead Lag control approach to design a power system stabilizer. An expression for synchronizing and damping torque coefficients with optimal controller is established. 1.2 SYSTEM INVESTIGATION A single machine-infinite bus (SMIB) system is considered for the present investigations. A machine connected to a large system through a transmission line may be reduced to a SMIB system, by using Thevenins equivalent of the transmission network external to the machine. Because of the relative size of the system to which the machine is supplying power, the dynamics associated with machine will cause virtually no change in the voltage and frequency of the Thevenins voltage (infinite bus voltage). The Thevenin equivalent impedance shall henceforth be referred to as equivalent impedance (i.e. Re+jXe). The synchronous machine is described as the fourth order model. The two-axis synchronous machine representation with a field circuit in the direct axis but with out damper windings is considered for the analysis. The equations describing the steady state operation of a synchronous generator connected to an infinite bus through an external reactance can be linearized about any particular operating point as follows: Tm P = M d 2 dt 2 (1) (2) (3)
' P = K1 + K 2 Eq
' Eq =
K3 K3 K 4 E fd ' 1 + sTd 0 K 3 1 + sTd' 0 K 3
(4) The K-constants are given in appendix. The interaction between the speed and voltage control equations of the machine is
' Vt = K 5 + K 6 Eq
expressed in terms of six constants k1-k6.[2] These constants with the exception of k3, which is only a function of the ratio of impedance, are dependent upon the actual real and reactive power loading as well as the excitation levels in the machine. The idea of supplementary excitation control known as power system stabilizer (PSS) is to apply a signal through the excitation system to increase the damping torque of the generator Because of the phase lags of (1+sTA) and (1+sTdo K3) for s = jn, the following conditions must be met by the PSS: (a) A phase lead compensation must be included in PSS design so as to have a damping torque TE in phase with at the oscillating frequency n. (b) A gain must be included in PSS design in order to have an adequate magnitude of damping. Generally, the block diagram of a single-input PSS is shown in Figure 1.1 below.
sT w 1 + sT w
Figure 1.1: general block diagram of PSS Various input signals can be used for PSS design. The common used signals are: (a) Speed deviation (b) Frequency deviation f (c) Electric power deviation Pe (d) Accelerating power Pa Conventional PSS comprising cascade connected lead networks with generator angular speed deviation (D) as input signal has been considered. Figure 1.2 shows the linearized model of the
single machine connected to large system (SMIB) around the operating point. From the transfer function block diagram, the following state variables are chosen for single machine system. The linearized differential equations can be written in the form state space form as X (t ) = Ax(t ) + Bu (t ) (5) Where
X ( t ) = Eq E fd 0 D M A= K4 ( K A TA ) K 5 314 K1 M 0 0
T
(6)
0 K2 M 1 K 3T ' do ( K A TA ) K6 0 1 T ' do -1 TA 0
(7)
B=
0 0 0 , x = E q KA E TA FD
(8)
y = [ , Vt ]T
(9)
System state matrix A is a function of the system parameters, which depend on operating conditions. Control matrix B depends on system parameters only. Control signal U is the PSS output. From the operating conditions and the corresponding parameters of the system considered, the system eigenvalues are obtained.
1.3
CONVENTIONAL PSS AND DESIGN CONSIDERATIONS The exciter considered here is only having the gain of KA and the time constant of TA. The typical PSS consists of a washout
function, a phase compensator (lead/lag functions), and a gain. It is well known that the performance of the PSS is mostly affected by the phase compensator and the gain. Therefore, these are the main focus of the tuning process. Two first order phase compensation blocks are considered. If the degree of compensation required is small, a single first-order block may be used. Generally slight under compensation is preferable so that the PSS does not contribute to the negative synchronizing torque component. Wash out function (Tw) has the value of anywhere in the range of 1 to 20 seconds. The main considerations are that it should be long enough to pass stabilizing signals at the frequencies of interest relatively unchanged, but not so long that it leads to undesirable generator voltage excursions as a result of stabilizer action during system- islanding conditions. For local mode of oscillations in the range of 0.8 to 2 Hz, a wash out of 1.5 sec is satisfactory. From the view point of low-frequency interarea oscillations, a wash out time constant of 10 seconds or higher is desirable, since low- time constants result in significant phase lead at low frequencies. The stabilizer gain K has an important effect on damping of rotor oscillations. The value of the gain is chosen by examining the effect for a wide range of values. The damping increases with an increase in stabilizer gain up to a certain point beyond which further increase in gain results in a decrease in damping. Ideally, the stabilizer gain should be set at a value corresponding to maximum damping. However the gain is often limited by other considerations. The transfer function model of the SMIB system with the PSS is given in Figure 1.3. The transfer function of Power System Stabilizer is given by
1 + sT1 1 + sT3 10s *( )( ) H ( s) = K* 1 (1 + 10s) 1 + T2 s 1 + T4 s
(10)
Where, K : Tw : T1-T4 : PSS gain washout time constant phase lead time constants
A low value of T2=T4=0.05 sec is chosen from the consideration of physical realization. Tw=10 sec is chosen in order to ensure that the phase shift and gain contributed by the wash out block for the range of oscillation frequencies normally encountered is negligible.
Figure 1.2 : Linearized model of SMIB system
Figure.1.3 : Linearized model of SMIB system
The wash out time constant (Tw) is to prevent steady state voltage
off sets as system frequency changes. Considering two identical cascades connected lead-lag networks for the PSS T1=T3 . Hence now the problem reduces to the tuning of gain (K) and T1 only. The parameters of the PSS obtained for the damping ratio of 0.3. The oscillation frequency is generally about 0.8-2 Hz for the local mode of oscillations. In this SMIB system only local mode of oscillations are considered for the tuning of PSS. The local mode of oscillation occurs when a machine supplies power to a load center over long, weak transmission lines.
1.3.1 Conventional PSS design The eigenvalues of the above A matrix are obtained using Matlab. It is evident from the open loop eigenvalues, the system without PSS is unstable and therefore it is necessary to stabilize the system by shifting these eigenvalues to the LHP and far off from the imaginary axis. The location of the desired eigenvalues is calculated by choosing a damping factor z for the dominant root. 2 The real part is -w and the imaginary part is w 1 Where
wn is the undamped natural frequency of the corresponding root. For the determination of PSS parameters a damping factor of =0.3 is chosen (maximum damping). Corresponding to this damping factor the desired eigenvalues are obtained as.
1 = n + n 1 2
2 = n n 1
2
(11) (12)
It is to be noted here that the some of the eigenvalues need not be shifted since they are placed for off in LHP. If any electromechanical modes of oscillations are present then PSS needs to be added to enhance the dynamic stability of the system. By using Decentralized modal control (DMC) algorithm the parameters of the conventional power system stabilizer are found.
1.4
PROPOSED LQR POWER SYSTEM STABILIZER ALGORITHM
Minimizing the error criteria in equation (14) the controller generates the parameters of the gain and the phase lead time constant. Consider a linear system characterized by equation (5) where (A, B) is stabilizable. We define the cost index then determine the matrix K of the LQR vector
u = Kx So in order to minimize the performance index
J ( x, u , Q, R ) = ( xT Qx + uT Ru )dt , Q 0,R>0 0
(13) (14)
Where Q and R are the positive-definite Hermitian or real symmetric matrix. Note that the second term on the right side account for the expenditure of the energy on the control efforts. The matrix Q and R determine the relative importance of the error and the expenditure of this energy. From the above equations we get
J = ( xT Qx + xT K T RKx)dt = xT (Qx + K T RK ) xdt 0 0 1/ 2 ) is detectable and ( A BK ) is stable. (A,Q
(15)
where The linear quadratic regulation problem is to find a control u = Kx law such that and J is minimized, the solution is given by K = R 1BT P (16) The controller K can be derived using parameter-optimization techniques, we set
d xT (Q + K T RK ) x = ( xT Px) dt
(17)
Then we obtain
xT (Q + K T RK ) x = xT Px xT Px
(18) Comparing both sides of the above equation and note that this equation must hold true for any x, we require that
= xT [( A BK )T P + P( A BK )]x
10
( A BK )T P + P( A BK )] = (Q + K T RK )
(19) (20)
R =T T Where T is a nonsingular matrix And

T
AT P + PA + [TK (T T )1 BT P ]T
(21) The minimization J of with respect to K requires the minimization of xT [TK (T T )1 BT P ]T [TK (T T )1 BT P ]x (22) Which this equation is nonnegative, the minimum occurs when it is zero, or when TK = (T T )1 BT P (23) Hence
K = T 1(T T )1 BT P = R 1BT P
[TK (T T )1 BT P ] PBR 1BT + Q = 0
(24) (25) (26)
Thus we get a control law as
u (t ) = Kx(t ) = R 1BT Px(t ) In which P must satisfy the reduced Riccati equation:
PA + AT P PBR 1BT P + Q = 0
1.4.1
Weight Matrix Selection
The LQR design selects the weight matrix Q and R such that the performances of the closed loop system can satisfy the desired requirements mentioned earlier.The selection of Q and R is weakly connected to the performance specifications, and a certain amount of trial and error is required with an interactive computer simulation before a satisfactory design results. Now given these linear models we can use LQR or pole placement techniques to design full state feedback controllers, u=-Kx. The lqr function allows you to choose two parameters, R and Q, which will balance the relative importance of the input and state in
Comparison of LQR and Lead Lag Controller for Tuning Power System 11 for Damping Oscillations
the cost function that you are trying to optimize. The simplest case is to assume R=1, and Q=C'*C. Essentially, the lqr method allows for the control of all outputs. In this case, it is pretty easy to do. The controller can be tuned by changing the nonzero elements in the Q matrix to get a desirable response the Matlab function lqr can be used to derive optimal control gains for a continuous controller.
1.4.2 Numerical example
Consider the mathematical linearize state space model which represents the power system in equations (5)-(9) Choosing the machine parameters and nominal operating point as;
Machine (p.u): ' xd = 1.70 , xd = 0.245 xq = 1.64 , T ' = 5.9 s, D= 0.0, and M=2H d0 = 4.74. Transmission line (p.u): re = 0.02, xe = 0.4 Exciter : KA = 400 TA= 0.05 s Operating point: Vt0 = Eb0 = 1.0 ,KE=0.17,TE=0.95 P0 = 1.0, Q0 = 0.1 and 0 = 53.74 0 The A matrix given in equation. (7) is evaluated as 377 0 0 0 1.423 0 1.9 0 A= 0.1 0 0.1 0.10 1984.2 0 8, 004.6 20 Consider the mechanical torque is constant and the vector B is obtained as follows: T B = [ 0 0 0 8, 000] The open-loop poles are: 1,2 = -13.6367 j30.0225
34 =33.5981 j25.8888
12
The dominant poles -13.6367 j30.0225 can be rewritten as:

1, 2 = n + j n 1 2
Where : = damping coefficient; n = frequency. The real part of dominant pole from equation (13) is -13.6367 and imaginary part is j30.024 which implies = 0.414 and n = 32.98rad / s f = 2.16 Hz It is observed that that the electromechnical mode characterized by pair of eigenvalues 3.6 j25.9 is negatively damped. The desired value of the damping coefficient to damp the oscillation of speed is obtained by optimum controller after shifting the unstable poles (3.6 j25.9 to -10.5j20.8) The dominant closed loop poles are now specified at =0.365 and
2 n=34.8radsThe imaginary part is now at n 1 = 27.7 The new system is now stable with LQR Controller in closed loop:
0.015
Light load DW response
0.01
s p e e d d e v ia tio n
0.005
-0.005
-0.01
Conventional PSS NO PSS LQR PSS

10 20 30
-0.015 0
Time in sampling of 0.1
40
50
60
70
80
90
100
Figure.1.4 Vs time for light load condition of the SMIB system

Normal load DW esponse
0.025 0.02 0.015
p ed d eviatio n
0.01 0.005 0 -0.005 -0.01 -0.015 -0.02 -0.025 0 20 40 60 80
N0 PSS Conventional PSS LQR PSS Time in 0.1 sampling

100 120
Figure. 1.5 Vs time for normal load condition of the SMIB system
Heavy Load Speed deviation DW response
0.04 0.03 0.02 NO PSS Conventional PSS LQR PSS
s p e e d d e v ia tio n
0.01 0
-0.01 -0.02 -0.03 -0.04 -0.05 0 10 20 30 40 50 60 70 80 90 100
Time in 0.1 sampling
Figure.1.6 Vs time for Heavy load condition of the SMIB system
14

0.015
Light load DW response
0.01
s p e ed d ev iatio n
0.005
-0.005
-0.01
Conventional PSS NO PSS LQR PSS

10 20 30
-0.015 0
Time in sampling of 0.1
40
50
60
70
80
90
100
Figure 1.7 Vs time for Light load condition of the SMIB system
Normal load angle deviation
4 3 2 1 0 -1 -2 -3 0 No PSS Conventional PSS LQR PSS
An g le d eviatio n
20
40
Time 0.1 sampling
60
80
100
120
Figure1..8 Vs time for Normal load condition of the SMIB system

Heavy load excitor voltage response
200 NO PSS Conventional PSS LQR PSS
150
E x c ito r v o lta g e (E fd )
100
50
-50
-100 0
20
40
Time in 0.1 Samping time
60
80
100
120
Figure 1.9 Efd Vs time for Heavy load condition of the SMIB system
Table 1.1 Operating conditions
Case 1 Normal 2 Heavy 3 Light
P(pu) 1.0 1.2 0.2
Q(pu) 0.1 0.2 0.05
Table 1.2 Eigenvalues
OPEN LOOP EIGENVALUE Normal - 13.6j30.0 3.6 j25.9 Normal Heavy -13.2j31.2 3.1j25.3 Light -10.7j27.7 0.67j24.6
CLOSED LOOP EIGENVALUE Heavy Light -12.835.7 -9.96 20.5 -9.629.0 -5.7 23.2
-12.7 j34.8 -10.5j20.8
Table 1.3 Controller parameters CONVENTIONAL PSS LQR Constants
7.7 0.05 0.05 0.2287 0.2287
K K1 K2 K3 K4
Normal 0.6292 0.8050 0.8413 0.0033
Heavy 0.5407 9.4519 0.8061 0.0032
Light 0.6315 5.5692 0.2576 0.0013
T2 T4 T1 T3
16
1.5 COMPARISON OF VARIOUS DESIGN TECHNIQUES
The linearized incremental state space model for a single machine system with its voltage regulator with four state variables has been developed. The single machine system without PSS is found unstable with roots in RHP. The system dynamic response without PSS is simulated using Simulink for 0.05 p.u disturbance in mechanical torque. MATLAB coding is used for conventional PSS design techniques, the final values of gain (K), and phase lead time constant (T) are given to the Simulink block. While LQR PSS, design techniques obtain a vector of K from Matlab function lqr are also given to the Simulink block. The dynamic response curves for the variables , and Efd are taken from the Simulink. The system responses curves of open loop system, the conventional PSS (CPSS), LQR based PSS are ompared. Shaft speed deviation is taken as the input to the all the Power system stabilizers. So the PSS is lso called as delta-omega PSS. The system dynamic response with PSS is simulated using these Simulink agrams for 0.05 p.u step change in mechanical torque Tm. The dynamic response curves for the variables change in speed deviation (), change in rotor angle deviation () and change in Excitor voltage deviation (Efd) of the single machine system with PSS are plotted for three different types of power system stabilizers (PSS) are shown in Figures 1.41.9. It is observed that the oscillations in the system output variables with PSS are well suppressed. Table 1-3 shows the eigenvalues of different load condition while the Tables 1.1-1.3 shows conventional PSS parameters and LQR parameters for different load condition
1.6 SIMULATION RESULTS
Performance of fixed-gain CPSS is better for particular operating conditions. It may not yield satisfactory results when there is a drastic change in the operating point.
The proposed PSS has robustness control property with power system operating points change and its parameters variation and uncertainty Dynamic response shows that the LQR based PSS has optimum response and the response is smooth and it has less over shoot and settling as compared to conventional PSS. The response curves shows that LQR PSS has less over shoot and settling time as compared to the open loop response and the traditional Lead-lag PSS.
1.7 CONCLUSION In this study, an optimal control algorithm is proposed to the PSS design problem. The method is illustrated by applying it to PSS controller design for damping oscillations. The procedure is developed identifying the most effective optimum controller based on state feedback. A numerical example illustrates the effectiveness of LQR damping performance by shifting all the negative damping poles to positive damping poles. The potential of the proposed design approach has been demonstrated by comparing the response curves of various power system stabilizer (PSS) design techniques. In addition, the simulation results show that the proposed LQR can work effectively and robustly over wide range of loading conditions and system configurations.
APPENDIX Derivation of K-constants All the variables with subscript 0 are values of variables evaluated at their pre-disturbance steady-state operating point from the known values of P0 , Q0 and Vt0. P0Vto iq 0 = (27) ( P0 x q ) 2 + (Vt 2 + Q0 x q ) 2 0
vd 0 = iq 0 xq
(28)
18
vqo = Vt 2 vt20 0
id 0 =
2 Q0 + xqiq 0
(29) (30) (31) (32) (33) Eq 0 E0 cos 0 xe + xq (34)
vq 0
Eq 0 = vq 0 + id 0 xq E0 = (vd 0 + xeiq 0 ) 2 + (vq 0 xe id 0 ) 2

(vd 0 + xeiq 0 ) (vq 0 xeid 0 )
' d
0 = tan
K1 = K2 =
' xq xd
xe + x
iq 0 E0 sin 0 +
E0 sin 0 ' xe + xd
' xd + xe xd + xe
(35) (36) (37)
K3 = K4 =
' xq xd ' xe + xd
E 0 sin 0
xq vd0 xd' vq0 K5 = E0 cos0 E0 sin0 ' xe + xq V0 xe + xd V0 t t

K6 = xe vq 0 ' xe + xd Vt 0
(38) (39)
Nomenclature All quantities are per unit on machine base. D Damping Torque Coefficient M Inertia constant Angular speed Rotor angle Id, Iq Direct and quadrature components of current
armature
Xd, Xq Synchronous reactance in d and q axis Xd, Xq Direct axis and Quadrature axis transient reactance EFD Equivalent excitation voltage KA Exciter gain TA Exciter time constant Tm,Te Mechanical and Electrical torque Tdo Field open circuit time constant. Vd,Vq Direct and quadrature components of terminal voltage K1 Change in Te for a change in with constant flux linkages in the d axis K2 Change in Te for a change in d axis flux linkages with constant K3 Impedance factor K4 Demagnetising effect of a change in rotor angle K5 Change in Vt with change in rotor angle for constant Eq K6 Change in Vt with change in Eq constant rotor angle
REFERENCES [1] Liqun Xing A Comparison of Pole Assignment & LQR Design Methods for Multivariable Control for STATCOM Master Theses, Department of Mechanical Engineering, College Of Engineering, Florida State University .2003. [2] J. Chow, J.Sanchez-Gasca, H. Ren and Sh. Wang,Power System damping Controller Design Using Multiple Input Signals, IEEE Control Systems Magazine, Volume 20, 82-90, August 2000 [3] Joe H.Chow, George E.Boukarim and Alexander Murdoch, Power system stabilizers as undergraduate
20
control design projects, IEEE trans. on power systems, vol 19, pp.144-151, Feb.2004. [4] Y.L.Abdel-Maidand M.M. Dawoud, Tuning of power system stabilizers using genetic algorithms, Electric Power Systems Research,Vol. 39, pp. 137-143, Jul.1996. [5] [M.E. Aboul-Ela, A.A. Salam, J.D. McCalley and A.A. Fouad, Damping Controller Design for Power System Oscillations Using Global Signals, IEEE Transactions on Power Systems, Vol. 11, No2, pp. 767-773, May 1996. [6] M.Klein, G.J.Rogers, S.Moorty, P.Kundur, Analytical investigation of factors influencing power system stabilizers performance, IEEE trans. On energy conversions, vol.7, No.3, pp. 382-390, Sep.1992. [7] Yao-nan, Qing-hua Li , Pole placement power system stabilizers design of an unstable nine- machine system, IEEE transactions on power systems, Vol 5, No.2, pp.353358,1990. [8] Bikash Pal , Balarko Chaudhuri Robust Control in power System power electronics and ower series Editors. New York USA Springer 2005 [9] L. Fan and A. Feliachi, Robust TCSC Control Design for Damping Inter-Area Oscillations, Power Engineering Society Summer Meeting, 2001. IEEE Vol. 2, pp.784 789., 15-19 July 2001. [10] Kazemi, and M. V. Sohforouzani Power System Damping Using Fuzzy Controlled FACTS Devices International Conference on Power System Technology POWERCON 2004 Singapore, 21-24 November 2004.
Eigenvalues Based Method for Analysis of Power System with SVC Device for DampingOscillations
21
2
EIGENVALUES BASED METHOD FOR ANALYSIS OF POWER SYSTEM WITH SVC DEVICE FOR DAMPING OSCILLATIONS
N. Magaji M.W. Mustafa
2.1 INTRODUCTION
The concept of flexible ac transmission systems (FACTS) is made possible by the application of high power electronic devices for power flow and voltage control [l]. In addition a number of FACTS devices have already been installed to aid power system dynamics which help to mitigation a low frequency oscillations often arise between areas in a large interconnected power network [2]. Eigenvalue sensitivities are one important outcome of the modal analysis and control of oscillatory behavior and dynamic stability in power systems. The pioneering work [3] considers the local oscillation of a single machine by means of a transfer function model. The usually complex pattern of oscillations in a large power system can be studied through linear, time invariant, statespace models based on the perturbations of the system state variables from their nominal values at a specific operating point Power system oscillations occur due to the lack of damping torque at the generators rotors. The oscillation of the generators rotors cause the oscillation of other power system variables (bus voltage, bus frequency, transmission lines active and reactive powers, etc.). Power system oscillations are usually in the range
22
between 0.1 and 2 Hz depending on the number of generators involved in [4]. Local oscillations lie in the upper part of that range and consist of the oscillation of a single generator or a group of generators against the rest of the system. In contrast, inter-area oscillations are in the lower part of the frequency range and comprise the oscillations among groups of generators. In addition, power system oscillations exhibit low damping compared to oscillations found in other dynamic systems: an oscillation of 10% damping is commonly accepted as well damped. To improve the damping of oscillations in power systems, supplementary control laws can be applied to existing devices. These supplementary actions are referred to as power oscillation damping (POD) control This work reviews the basic concepts of eigenvalues analysis of linear systems. The physical meaning of eigenvalues, eigenvectors, participation factors, residues and controllability and observability indices will be introduced and illustrated in small scale power systems. This technique has been successfully used in location and tuning of power system stabilizers [5] The application of sensitivity measures to the design of power system damping (POD) controllers has been applied to SVC. The design method utilizes the residue approach; this presented approach solves the optimal sitting of the SVC device, selection of the proper feedback signals and the controller design problem
2.2 BASIC CONCEPTS OF LINEAR SYSTEMS ANALYSIS
Low frequency electromechanical oscillations range from less than 1 Hz to 2 Hz other than those with sub-synchronous resonance (SSR) [6]. Multi-machine power system dynamic behavior in this frequency range is usually expressed as a set of non-linear differential and algebraic (DAE) equations. The algebraic equations result from the network power balance and generator stator current equations. The initial operating state of the algebraic variables such as bus voltages and angles are obtained through a
23
standard power flow solution. The initial values of the dynamic variables are obtained by solving the differential equations
2.2.1 Eigenvalues, Eigenvectors and Modes Let us start from the mathematical model a dynamic system expressed in terms of a system of non-linear differential equations: x = F ( x, t ) (1) If this system of non-linear differential equations is Linearized around an operating point of interest x=x0, it results in: x = Ax(t ) (2) A meaningful solution method of equation (2) is based on the eigenvalues and eigenvectors of the state matrix A. An eigenvalue i of the state matrix A and the associated right vi and left wi eigenvectors are defined according to: Avi = i vi In a matrix with all distinct eigenvalues (not a necessity but it is easier to understand when it is so), all the right eigenvectors and eigenvalues can be expressed as a compact matrix expression AV = VA (3) where,
V = ( v 1 ,v 2 ..... v n 1 v n , )
Pre-multiplying both sides of (3) by V-1 gives

V 1 AV =
= diag ( 1 2 ... n-1 , n )
(4) (5) (6) (7)

-l
A similar expression holds for the left eigenvectors W such that

WA = W
Where
t t t W = [ 1t , 2 ..... n 1 n ]t
Post-multiplying both sides of (6) by W , gives (8) The transformed physical state variables (x) can be put into modal variables of equation (2) with the help of eigenvector matrices V and W x = Vz (9)
WAW 1 =
z = Wx
24
In power system literature, the right eigenvector matrix v is known as the mode shape matrix, that is, eigenvector vi is known as the ith mode shape, corresponding to eigenvalue i. The mode shape provides important information on the participation of an individual machine or a group of machines in one particular mode. solution of equation (2) can be expressed in terms of the eigenvalues and eigenvectors of the state matrix as:
x(t) = Ve t Wx(0) = vi ei t [w iT x(0)]
i =1 N
(10)
The analysis of equation (10) allows drawing the following conclusions: The system response is the combination of the system response to each of the N modes. The eigenvalues determine the system stability. A real positive (negative) eigenvalue determines exponentially increasing (decreasing) behavior. A complex eigenvalue of positive (negative) real part results in a increasing (decreasing) oscillatory behavior. The components of the right eigenvector vi measure the relative activity of each variable in the ith mode. The components of the left eigenvector wi weight the initial conditions in the i-th mode For the Participation factors it is natural to suggest that the significant state variables influencing a particular mode are those having large entries corresponding to the right eigenvector of i. The participation factor of the j-th variable in the k-th mode is defined as the product of the j-th's components of the right vjk and left wki eigenvectors corresponding to the k-th mode [7] Pjk =W jkVkj (11) The product W jkVkj is a dimensionless measure which is called participation factor. In other words, they are independent on the units of the state variables. In addition, both the sum of the participation factors of all variables in a mode and the sum of the participation of all modes in a variable are equal to one. Other interesting measure is the subsystem participation. The subsystem
25
participation is the magnitude of the sum of the participation factors of the variables that describe a subsystem in a mode.
2.2.2
Modal controllability and observability factors
The effectiveness of control in power system can be indicated through controllability and observability indices. This is important as control cost is influenced to a great deal by the controllability and observability of the plant. These issues are addressed through modal controllability and observability
2.2.3 Controllability index
Assume that an input u(t) and an output y(t) of the linear dynamic system equation (2) have defined
x(t) = Ax(t) + Bu(t) y(t) = Cx(t)
(12)
The application of a linear transformation defined by the eigenvectors of the state matrix to the system as described by (12) results in: equation (13): Let v and w be the right and left eigen vector matrices of A, respectively. If eigenvalues of A are distinct, then wTv = I, where wT is conjugate transpose of w and I is the identity matrix. Substituting x =wz in (12), we obtain
z(t) = w T Awz(t) + w T Bu(t) y(t) = cwz(t)
(13)
Equation (13) can be written for kth eigen mode as

z( t ) = k z k ( t ) + w T Bi vi ( t ) k
i =1 m
(14)
Where wk is the left eigenvector corresponding to kth mode and Bi is the ith column vector of matrix B. From (14), one can find the controllability of kth eigen mode with respect to the ith input. The
26
controllability index (CI) of an ith input to the kth mode [8,14] is defined as CIi = w T Bi (15) k T The input i, for which the value of wk Bi is maximum, is considered the suitable parameter to be controlled for affecting the kth eigen mode to maximum extent.
2.2.4 Observability index
The observability index (cvi) of an ith input to the kth mode is defined as (16) The study of equations (15) and (16) leads to the following conclusions: CIi Measures the controllability of the mode associated to the variable x i ( t ) from the input u( t ) .In other words, if the mode i can be controlled from the input u( t ) OIi Measures the observability of the mode associated to the variable x i ( t ) form the output y( t ) . In other words, if the mode i can be observed from the variable y( t ) Therefore, a mode can be controlled if only if it is controllable from the input u( t ) and observable from the output y( t )
OIi = Ci wk
2.2.5 Residue Considering (12) with single input and single output (SISO) and assuming D = 0, the open loop transfer function of the system can be obtained by
G( s ) =
y( s ) u( s )
(17)
= C( sI A )1 B
27
The transfer function G(s) can be expanded in partial fractions of the Laplace transform of y in terms of C and B matrices and the right and left eigenvectors as
Yref ( s )
G(s)
Y ( s)
H (s)
Figure. 2.1 Closed loop system with POD controller
G( s ) = =
N
C i B i i =1 ( s i )
N
Ri ( s i ) i =1
(18)
Each term in the denominator, Ri, of the summation is a scalar called residue. The residue Ri of a particular mode i gives the measure of that modes sensitivity to a feedback between the output y and the input u; it is the product of the modes observability and controllability. Figure 2.1 shows a system G(s) equipped with a feedback control H(s). When applying the feedback control, eigenvalues of the initial system G(s) are changed. It can be proven, that when the feedback control is applied, the shift of an eigenvalues can be calculated by i = Ri H( i ) (19)
28
It can be observed from (19) that the shift of the eigenvalue caused by the controller is proportional to the magnitude of the corresponding residue. For a certain mode, the same type of feedback controls H(s), regardless of its structure and parameters can be tested at different locations. For the mode of the interest, residues at all locations have to be calculated. The largest residue then indicates the most effective location to apply the feedback control.
2.2.6 Power Oscillation Damper Controller Design
Supplementary control action applied to FACTS devices to increase the system damping is called Power Oscillation Damping (POD). Since FACTS controllers are located in transmission systems, local input signals are always preferred, usually the active or reactive power flow through FACTS device or FACTS terminal voltages. Figure 2.1 shows the considered closed-loop system where G(s) represents the power system including FACTS devices and H(s) FACTS POD controller. In order to shift the real component of i to the left, FACTS POD controller is employed. That movement can be achieved with a transfer function consisting of an amplification block, a wash-out block and mc stages of leadlag blocks. We adapt the structure of POD controller given in [8, 12], i.e. the transfer function of the FACTS POD controller is:
sTw 1 H (s) = K * * 1 + sTm 1 + sTw = KH 1 ( s ) 1 + sTlead 1 + sT lag
mc
(20) Where K is a positive constant gain, and H1(s) is the transfer function of the wash-out and lead-lag blocks. The washout time constant, Tw, is usually equal to 5-10 s. The lead lag parameters can be determined using the following equations: co m p = 1 8 0 0 a rg ( R i ) (21)
29
1-sin comp mc Tlead c = = Tlag comp 1+sin mc 1 T la g = , T le a d = c T la g i c
(22)
(23)
Where arg(Ri) denotes phase angle of the residue Ri, i is the frequency of the mode of oscillation in rad/=sec, mC is the number of compensation stages (usually mC = 2),the angle compensated by each block should be 30-50. The controller gain K is computed as a function of the desired eigenvalue location ides according to equation 24:
K=
i d . Ri H1 (i )
(24)
2.3 SVC DEVICES The Static VAr Compensator (SVC) is a shunt connected device whose main functionality is to regulate the voltage at a chosen bus by suitable control of its equivalent reactance. The model that is used here assumes a time constant regulator, as depicted in Figure 2.2. In this model, a total reactance bSVC is assumed and the following differential equation holds [13]. bSVC = (K r (Vref + vPOD - V ) - bSVC )/Tr (25)
The model is completed by the algebraic equation expressing the reactive power injected at the SVC node: Q = -bSVCV 2 (26) The regulator has an anti-windup limiter, thus the reactance bSVC is locked if one of its limits is reached and the first derivative is set to zero.
30
v POD
bmax
Kr Tr s + 1
bsvc
Vref
bmin
Figure 2.3 SVC Topology
Figure 2.2 SVC Regulator
2.4
EIGENVALUE ANALYSIS OF POWER SYSTEMS
The concepts detailed in the previous section will be illustrated considering two small-scale power systems. The size of these systems allows the computation of all eigenvalues and eigenvectors of the state matrix without employing advance techniques due to small sizes of the system.
2.4.1 Analysis of single machine connected to an infinite bus with SVC
The case of a single generator connected to an infinite bus is considered first with and without SVC. The generator model contains accurate representations of the synchronous machine, the excitation and the speed-governing systems. It has been assumed that the generator is equipped with a static excitation system [11]. A static var compensator is connected at bus 2. The linear model of this system is described by 11 state variables. The synchronous machine, the exciter and SVC are described respectively by 6, 4 and 1 state variables. The eigenstructure of the state matrix
31
contains 3 pairs of complex eigenvalues and 5 real eigenvalues which are detailed in tables 2.1 and 2.2 respectively. Eigenvalues accurately determine linear systems stability: this system is close to instability due to the presence of a poorly damped oscillatory mode. However, if the connections between eigenvalues and state variables are sought, participation factors have to be used. Table 2.3 details the participation of the generator subsystems: rotor dynamics, synchronous machine, exciter, and SVC in all modes. Table 2.2 clearly indicates that the poorly damped oscillatory mode (eigenvalues 1 and 2) is associated to the rotor dynamics and that the other oscillatory mode (eigenvalues 3 and 4) describes the interaction between the synchronous machine and the exciter. The mode associated to the rotor dynamics is also known as electromechanical mode. Table 2.3 also shows that three exponential modes are associated to the machine (damper windings), other three to Exciter and the remaining mode to the SVC. The slower modes correspond to the SVC dynamics whereas the fastest mode is associated to the exciter
32
Table 2.1 Complex eigenvalues of SMIB with SVC
No. 1,2 3,4 5,6 1,2
Eigenvalue Frequency (Hz) Damping % 0.298357.8548 1.2501 3.80 0.047180.20797 0.0331 22.12 1.4854 14.1564 2.2531 10.44 Complex eigenvalues of SMIB without SVC 0.29835j7.8548 1.222 2.57
Table 2.2 Real eigenvalues OF SMIB with SVC No. 7 8 9 10 11 Eigenvalue Time constant -2.0542 -0.487 -36.7814 -0.027 -24.2903 -0.041 -66.6694 -0.015 -1.8333 -0.545
Table 2.3 Residue corresponding to local mode -0.29835+j7.8548
Transfer function Residue Phase angle -2.5350 + 1.0789i -23.05470 Pg

V
r
0.0010 + 0.0022i
r
65.5560
V1
-0.2589 + 7.4683i -35.3803

r
33
2.4.2
Analysis of two areas four machine with SVC
In this study, a two area interconnected four machine power system shown in Figure 2.5 is considered. The system consists of four machines arranged in two areas inter-connected by a weak tie line [11]. The location of SVC is indicated in the diagram. The system is operating with area 1 exporting 400 MW to area 2. The equivalent circuit of the system is shown in Figure 2.5. Network and generator data can be found in [11]. We have assumed that the generators are equipped with static excitation systems and steam turbines. In this study, the SVC is treated as a variable capacitance. From table 2.4 three pairs of poorly damped eigenvalues are found. They result to be associated to the rotor dynamics. The slowest eigenvalues are associated to the speed-governing systems whereas the fastest are associated to the excitation
34
j0.15
2200MVA
j0.5
Et P Q
j0.93
EB
Figure.2.5 11 bus two area four machines system
35
Table 2.4 Eigenvector and normalized participation factor corresponding poor mode -0.29835+j7.8548
S/N 1 2 3 4 5 6 7 8 9 10 11
Right eigenvector 0.7492 -0.0006 + j0.0178 -0.0020 + 0.0264 0.0033 - j0.0071 -0.0753 + j0.0385 0.0205 - j0.0103 -0.0613 + j0.0161i 0.3732 + j0.5283 0.0046 + j0.0105 -0.0750 j0.0425 0
Left eigenvector -3.4830 + j0.0000 -0.0065 j0.0000 7.5335 + j0.0000 -2.0275 + j1.8330 -2.0275 j1.8330 -0.0060 j0.0126 -0.0060 + j0.0126 -0.1162 j0.0000 0.0130 + j0.0014 0.0130 j0.0014 0
Participation factor 0.48018
Participation state Machine angle
0.48018 0.02047 0.00072 0.01593 0.00119 0.0001 0.00037 0.00015 0.00037 0.00037
Machine speed 1
q-axis damper e q
d-axis damper e d q-axis damper e q d-axis damper e d Exiter v m Exiter v r1 Exiter v r2 Exiter v f
SVC
36
Table 2.4 details the participation of the generators in each mode. The fastest mode is associated with local modes: mode -0.7647 j7.5680 is dominated by generator G3 and G4 and mode -0.7514 7.3036i is dominated by generator G1 and G2. While mode 0.1211 3.7559i is an inter-area mode.
2.4.3 Design of SVC POD controller using residue method fortwo area test system
The uncontrolled system, Figure 2.5, has one inter-area oscillatory mode characterized by = -0.1211 j3.7559 with damping ratio = 3.22%. According to Table 2.7, the bus 8 has the largest residue and therefore the most effective location of the SVC and to apply the feedback control. Using the method presented above, POD controller parameters are calculated in order to shift the real part of the oscillatory mode, to the left half complex plane. The obtained transfer function for the SVC POD controller
H ( s) = K * 1 10 s 1 + 0.0674 s 1 + 0.0674 s * * 1 + 0.1s 1 + 10 s 1 + 1.044 s 1 + 1.044 s
Eigenvalue of our interest moves form the original location = 0.1211 j3.7559 to the desired location d= -0.38 j3.76 to give about 10% damping as:
K=
i d = 821 Ri H1 (i )
37
Table 2.5 Eigenvalue and participation factors of 11 bus with SVC

COMPLEX EIGENVALUE OF TWO AREA FOUR MACHINE TEST Mode Complex Frequency Damping Dominant state No. Eigenvalue (Hz) ratio % 1,2 3,4 5,6 7,8 9,10 11,12 13,14 15,16 17,18 19,20 21,22 23,24 25,26 -38.9130 j0.5195 -38.6438 j0.5235 -16.2015 j3.3165 - 16.1676 j2.7048 -0.7647 j7.568 -0.7514 j7.303 -0.1211 j3.756 -8.3489 j0.621 -8.3031 j0.5799 -0.8417 j0.4294 -0.5219 j0.4762 -0.3704 j0.3102 -0.3577 j0.3073 0.08 0.08 0.53 0.43 1.20 1.16 0.60 0.10 0.09 0.07 0.08 0.05 0.05 99.99 99.99 97.97 -98.63 10.05 10.23 3.22 99.72 99.76 89.08 73.87 76.67 75.85
e and e d q of G2,G4
e of G2 and d e of ,G1 q
REAL
EIGENVALUE OF TWO AREA FOUR MACHINE TEST
Real eigenvalue -70.3594 -34.2154 -32.9547 -24.2634 -22.9037 -17.906 -17.8612 -13.5968 -13.4863 -8.0211 -8.1299 -4.3305 -2.7976 0.0014 -0.0014
Associate state
BSVC
e of G1 q
Vr1 of Exc.3 and Vr1of Exc.4 Vr1 of Exc.2 and Vr1of Exc.1
, of G2 ,G3 and G4
, of G 1 ,G 3 a n d G 2
of G eq 4
of G ed 3 e of G 1 d
vm of Exc.4 vm of Exc.1 vr1 of Exc.1 vr1 of Exc.3 v f of Exc.2 v f of Exc.3
, of G3 ,G1 G4 and G2
Vf of Exc.3 and Vf of Exc.4 Vf of Exc.1 and Vf of Exc.2 e of G2 q and G1 e of G3 and q

Vr2 of Exc.4
e of G2 q and G1 e of G4 and q
Vr2 of Exc.3
e of G3 d e of G1 d
,of G 2 , of G2
38
Table 2.6
Mode No. 9,10 11,12 13,14
Participation of the generators in the electromechnical modes of the two area test system
Eigenvalue -0.7647 7.5680i -0.7514 7.3036i -0.1211 3.7559i G1 0.01145 0.41104 0.24613 G2 0.0429 0.5364 0.1420 G3 0.4101 0.0277 0.3459 G4 0.55216 0.01232 0.24344
Table 2.7 Sitting Indices of SVC for Two Area Four Machine Test
SVCBus location 5 6 7 8 9 10 11
Complex Residue 0.00023-0.00015 1.610-6 -6.6i10-7 (-8.2+2.53i)10-7 -0.0028 - 0.0043i (3.84-7.3i) 10-8 0.0031-0.0029i -0.00042 - 0.00063i
|Ri| 0.00028 1.7310-6 8.5810-7 0.0052 8.2311110-8 0.0042 0.00075
Table 2.8 Inter Area Mode Comparison for Test System 2
Mode No Operating Conditions Frequency (Hz) Damping ratio % 13,14 Without SVC 0.49 2.98 13,14 With SVC 0.6 3.22 13,14 SVC with POD 0.5913 20
39
2.5
CONCLUSIONS
This work has reviewed methods for analysis and control of power system oscillations with SVC device based on the Eigen structure of the state matrix of the linear model of the power system. Residue-based methods also provide valuable information on how to design power system damping controllers. Although eigenvalues based methods are very powerful, the complexity of the power system stability problem requires the complementary use of other methods such as non-linear time domain simulation. All the simulations were done with PSAT toolbox in Matlab environment.
REFERENCES [1] B. Kalyan Kumar, S.N. Singh and S.C. SrivastavaPlacement of SVC controllers using modal controllability indices to damp out power system oscillations IET Gener. Transmission . Distrib., Vol. 1, No. 2, page 209-217, March 2007 [2] RJ. Piwko, Ed., Applications of Static Var Systems for System Dynamic Performance, IEEE Publication 87TH01875-5-PWR, [3] N.G. Hingorani, Flexible ac transmission, lEEE Spectrum, April 993, pp. 40-45. [4] M. A. Pai and Alex Stankovic Robust control in power system tex book. USA, Springer Science+Business Media, Inc. 2005 [5] S.N. Singh, A.K. David A New Approach for Placement of FACTS Devices in Open Power Markets IEEE Power Engineering Review, September 2001pp 58-60 [6] K.R. Padiyar Power system dynamics stability and Control Anshan limited UK 2004 [7] E. Acha, C. Fuerte-Esquivel, H. Ambriz-Perez and C. Angeles- Camacho. FACTS Modelling and Simulation in Power Networks, John Wiley & Sons LTD, England, 2004.
40
[8] [9]
[10] [11] [12]
[13] [14]
Bikash Pal , Balarko Chaudhuri Robust Control in power System Springer Science+Business Media, Inc.USA 2005 kamoto, H., Kurita, A., and Sekine, Y.: A method for identification of effective locations of variable impedance apparatus on enhancement of steady state stability in large scale power systems, IEEE Trans. Power Syst., 1995, 10, (3), pp. 14011407 Y.H. Song and A.T. Johns. Flexible AC Transmission Systems. IEE Power and energy series, UK, 1999. Kundur P. Power system stability and control. New York, USA: McGraw- Hill; 1994. J. Perez-Arriaga, G.C. Verghese, F.C. Schweppe, Selective Modal Analysis with Applications to Electric Power Systems. Part I:Heuristic Introduction. Pant 11: The Dynamic Stability Problem. IEEE Transactions on Power Apparatus andsystems, Vol. PAS-101, No. 9, September 1982, pp. 3117-3134 Federico Milano, Documentation for Power System Analysis Toolbox PSAT version 2.0.0 , March 8, 2007. http://thunderbox.uwaterloo.ca/~fmilano Einar, V.L., and Joe, H.C.: Concepts for design of SVC controllers to damp power swings, IEEE Trans. Power Syst., 1995, 10, (2), pp. 948954
Damping controller Design for SVC Device using Residue Factor
41
3
DAMPING CONTROLLER DESIGN FOR SVC DEVICE USING RESIDUE FACTOR
M.W. Mustafa N. Magaji
3.1 INTRODUCTION
The concept of flexible ac transmission systems (SVC) is made possible by the application of high power electronic devices for power flow and voltage control [l]. In addition a number of SVC devices have already been installed to aid power system dynamics which help to mitigation a low frequency oscillations often arise between areas in a large interconnected power network [2]. Eigenvalues sensitivities are one important outcome of the modal analysis and control of oscillatory behavior and dynamic stability in power systems. The pioneering work [3] considers the local oscillation of a single machine by means of a transfer function model. The usually complex pattern of oscillations in a large power system can be studied through linear, time invariant, statespace models based on the perturbations of the system state variables from their nominal values at a specific operating point. Power system oscillations occur due to the lack of damping torque at the generators rotors. The oscillation of the generators rotors cause the oscillation of other power system variables (bus voltage, bus frequency, transmission lines active and reactive powers, etc.). Power system oscillations are usually in the range between 0.1 and 2 Hz depending on the number of generators involved in [4]. Local oscillations lie in the upper part of that range and consist of the oscillation of a single generator or a group of generators against
42
the rest of the system. In contrast, inter-area oscillations are in the lower part of the frequency range and comprise the oscillations among groups of generators. In addition, power system oscillations exhibit low damping compared to oscillations found in other dynamic systems: an oscillation of 10% damping is commonly accepted as well damped. To improve the damping of oscillations in power systems, supplementary control laws can be applied to existing devices. These supplementary actions are referred to as power oscillation damping (POD) control. This work reviews the basic concepts of eigenvalue analysis of linear systems. The physical meaning of eigenvalues, eigenvectors, participation factors, residues and controllability and observability indices will be introduced and illustrated in small scale power systems. This technique has been successfully used in location and tuning of power system stabilizers. In this work the application of sensitivity measures to the design of power system damping (POD) controllers has been applied to SVC. The design method utilizes the residue approach; this presented approach solves the optimal sitting of the SVC device, selection of the proper feedback signals and the controller design problem. The effectiveness of the approach is tested by computing the damping of no SVC, with SVC and SVC with POD. In addition a PI controller is used as a bench mark for testing the effectiveness of the controller.
LINEAR SYSTEM ANALYSIS TOOLS IN POWER SYSTEMS Low frequency electromechanical oscillations range from less than 1 Hz to 2 Hz other than those with sub-synchronous resonance (SSR) [5, 6]. Multi-machine power system dynamic behavior in this frequency range is usually expressed as a set of non-linear differential and algebraic (DAE) equations. The algebraic equations result from the network power balance and generator stator current equations. The initial operating state of the algebraic variables such as bus voltages and angles are obtained through a 3.2
43
standard power flow solution. The initial values of the dynamic variables are obtained by solving the differential equations
3.1.1 Eigenvalues, Eigenvectors and Modes Let us start from the mathematical model a dynamic system expressed in terms of a system of non-linear differential equations: x = F ( x, t ) (1) If this system of non-linear differential equations is Linearized around an operating point of interest x=x0, it results in: x = Ax(t ) (2) A meaningful solution method of (2) is based on the eigenvalues and eigenvectors of the state matrix A. An eigenvalue i of the state matrix A and the associated right vi and left wi eigenvectors are defined according to: Avi = i vi In a matrix with all distinct eigenvalues (not a necessity but it is easier to understand when it is so), all the right eigenvectors and eigenvalues can be expressed as a compact matrix expression AV = VA (3) where, V = ( v1 ,v2 ..... vn 1 vn , ) (4) = diag ( 1 2 ...n-1 ,n ) Pre-multiplying both sides of (3) by V-1 gives (5) V 1 AV = A similar expression holds for the left eigenvectors W such that WA = W (6) Where t t t W = [ 1t ,2 ..... n 1 n ]t (7) -l Post-multiplying both sides of (6) by W , gives WAW 1 = (8) The transformed physical state variables (x) can be put into modal variables (2) with the help of eigenvector matrices V and W
44
x = Vz (9) z = Wx In power system literature, the right eigenvector matrix v is known as the mode shape matrix, that is, eigenvector vi is known as the ith mode shape, corresponding to eigenvalue i. The mode shape provides important information on the participation of an individual machine or a group of machines in one particular mode. Solution of (2) can be expressed in terms of the eigenvalues and eigenvectors of the state matrix as:
x(t) = Vet Wx(0) = vi ei t [w iT x(0)]
i =1 N
(10)
3.2.2 Participation factors It is natural to suggest that the significant state variables influencing a particular mode are those having large entries corresponding to the right eigenvector of i. The participation factor of the j-th variable in the k-th mode is defined as the product of the j-th's components of the right vjk and left wki eigenvectors corresponding to the k-th mode [7] Pjk =W jkVkj (11) The product W jkVkj is a dimensionless measure which is called participation factor. In other words, they are independent on the units of the state variables. In addition, both the sum of the participation factors of all variables in a mode and the sum of the participation of all modes in a variable are equal to one. Other interesting measure is the subsystem participation. The subsystem participation is the magnitude of the sum of the participation factors of the variables that describe a subsystem in a mode. 3.2.3 Modal controllability and observability factors The effectiveness of control in power system can be indicated through controllability and observability indices. This is important as control cost is influenced to a great deal by the controllability
45
and observability of the plant. These issues are addressed through modal controllability and observability.
3.2.4 Controllability index Assume that an input u(t) and an output y(t) of the linear dynamic system (2) have defined
(12) y(t) = Cx(t) The application of a linear transformation defined by the eigenvectors of the state matrix to the system as described by (12) results in: equation (13): Let v and w be the right and left eigen vector matrices of A, respectively. If eigenvalues of A are distinct, then wTv = I, where wT is conjugate transpose of w and I is the identity matrix. Substituting x =wz in (12), we obtain
x(t) = Ax(t) + Bu(t)
(13)

z( t ) = k z k ( t ) + w T Bi vi ( t ) k
i =1 m
(14)
Where wk is the left eigenvector corresponding to kth mode and Bi is the ith column vector of matrix B. From (14), one can find the controllability of kth eigen mode with respect to the ith input. The controllability index (CI) of an ith input to the kth mode is defined as CIi = w T Bi (15) k T The input i, for which the value of wk Bi is maximum, is considered the suitable parameter to be controlled for affecting the kth eigen mode to maximum extent.
46
3.2.5 Observability index The observability index (cvi) of an ith input to the kth mode is defined as OIi = Ci wk (16)
3.3 Residues Considering (12) with single input and single output (SISO) and assuming D = 0, the open loop transfer function of the system can be obtained by
G( s ) = y( s ) u( s )
1
(17)
= C( sI A ) B
G( s ) = =
N
C i B i i =1 ( s i )
N
Ri ( s i ) i =1
(18)
Each term in the denominator, Ri, of the summation is a scalar called residue. The residue Ri of a particular mode i gives the measure of that modes sensitivity to a feedback between the output y and the input u; it is the product of the modes observability and controllability. Figure 3.1 shows a system G(s) equipped with a feedback control H(s). When applying the feedback control, eigenvalues of the initial system G(s) are changed. It can be proven, that when the feedback control is applied, the shift of an eigenvalues can be calculated by i = Ri H( i ) (19) It can be observed from (19) that the shift of the eigenvalue caused by the controller is proportional to the magnitude of the corresponding residue. For a certain mode, the same type of feedback controls H(s), regardless of its structure and parameters can be tested at different locations. For the mode of the interest, residues at all locations have to be calculated. The largest residue
47
then indicates the most effective location to apply the feedback control.
3.3 SVC MODEL The Static VAr Compensator (SVC) is a shunt connected device whose main functionality is to regulate the voltage at a chosen bus by suitable control of its equivalent reactance. A typical topology of a SVC comprises a parallel combination of a thyristor controlled reactor and a fixed capacitor as shown in Figure 3.1. A small signal dynamic model of SVC that is used here assumes a time constant regulator, as depicted in Figure 3.2. In this model, a total reactance bSVC is assumed and the following differential equation holds bSVC = (K r (Vref + vPOD - V ) - bSVC )/Tr (20)
The model is completed by the algebraic equation expressing the reactive power injected at the SVC node: Q = -bSVCV 2 (21) The regulator has an anti-windup limiter, thus the reactance bSVC is locked if one of its limits is reached and the first derivative is set to zero [8]. The supplementary input POD is used to connect the POD controller for damping oscillation while Vref to maintain acceptable voltage at the SVC bus.TCR of 150MVAr is connected in parallel with fixed capacitor of 200MVAr correspond to a limit of 2.0pu to -1.5 pu at 1.0 pu voltage [9].
48
TCR
Figure 3.1SVC Topology .
v POD
V
+
Kr Tr s + 1
bmax
bsvc
+
Vref
bmin
3.3.1 SVC POD Controller Design Supplementary control action applied to SVC devices to increase the system damping is called Power Oscillation Damping (POD). Since SVC controllers are located in transmission systems, local input signals are always preferred, usually the active or reactive power flow through SVC device or SVC terminal voltages. Figure
49
3.3 shows the considered closed-loop system where G(s) represents the power system including SVC devices and H(s) SVC POD controller.
Yref ( s )
G(s)
Y ( s)
H SVC ( s )
Fig. 3.3 Closed loop system with POD controller
In order to shift the real component of i to the left, SVC POD controller is employed. That movement can be achieved with a transfer function consisting of an amplification block, a wash-out block and mc stages of lead-lag blocks. We adapt the structure of POD controller given in [10, 8] i.e. the transfer function of the SVC POD controller is:
H (s) = K * sT w 1 * 1 + sT m 1 + sT w 1 + sT lea d 1 + sT la g
mc
(22) Where K is a positive constant gain and H1 is the transfer function of the wash-out and lead-lag blocks. The washout time constant, Tw, is usually equal to 5-10 s. The lead lag parameters can be determined using the equation (22)
= KH 1(s)
comp = 1800 arg( Ri )
(23)
50
c =
Tlead Tlag
1-sin comp mc = 1+sin comp mc
(24)
Tlag =
i c
, Tlead = c Tlag
25)
Where arg(Ri) denotes phase angle of the residue Ri, i is the frequency of the mode of oscillation in rad/=sec, mC is the number of compensation stages (usually mC = 2),the angle compensated by each block should be 30-50. The controller gain K is computed as a function of the desired eigenvalue location ides according to equation 26:
K=
i d . Ri H1 (i )
(26)
3.4 TEST MODELS 3.4.1 Two areas four machine with SVC In this study, a two area interconnected four machine power system shown in Figure 3.4 is considered. The system consists of four machines arranged in two areas inter-connected by a weak tie line [11].
51
Damping Controller
I 910
I ref
The location of SVC is indicated in the diagram. The system is operating with area 1 exporting 400 MW to area 2. The equivalent circuit of the system is shown in Figure 3.4. Network and generator data can be found in [11]. We have assumed that the generators are equipped with static excitation systems and steam turbines. In this study, the SVC is treated as a variable capacitance. Three pairs of poorly damped eigenvalues are found. They result to be associated to the rotor dynamics. The slowest eigenvalues are associated to the speed-governing systems whereas the fastest are associated to the excitation systems. The synchronous machine modes are in between. From the table 3.1 and 3.2, we see that the system is stable there are four rotor angle modes. There mode shapes are described by the component of the right eigenvector corresponding to the generator speed not shown.
52
Table 3.1 Participation of the generators in the electromechnical Modes of the two area test system
Mode No. Eigenvalue 9,10 -0.76477.5680i 11,12 -0.7514 7.3036i 13,14 -0.1211 3.7559i
G1 0.0115 0.4110 0.2461
G2 0.0430 0.5364 0.1420
G3 0.4101 0.0277 0.3459
G4 0.5522 0.0123 0.2434
DESIGN OF SVC POD CONTROLLER USING RESIDUE METHOD The uncontrolled system, Figure 3.4, has one inter-area oscillatory mode characterized by = -0.1211 j3.7559 with damping ratio = 3.22%. According to residue method, bus 8 has the largest residue and therefore the most effective location of the SVC and to apply the feedback control. Using the method presented above, POD controller parameters are calculated in order to shift the real part of the oscillatory mode, to the left half complex plane. The obtained transfer function for the SVC POD controller
H ( s) = K * 1 10 s 1 + 0.0674 s 1 + 0.0674 s * * 1 + 0.1s 1 + 10 s 1 + 1.044 s 1 + 1.044 s
3.5
K=
i d = 821 Ri H1 (i )
53
Table 3.1 Complex Eigenvalues of Two Area Four Machine Test
Mode No. 1,2 3,4 5,6 7,8 9,10 11,12 13,14 15,16 17,18 19,20 21,22 23,24 25,26
Complex Eigenvalue -38.9130 j0.5195 -38.6438 j0.5235 -16.2015 j3.3165 -16.1676 j2.7048 -0.7647 j7.5680 -0.7514 j7.3036 -0.1211 j3.7559 -8.3489 j0.6210 -8.3031 j0.5799 -0.8417 j0.4294 -0.5219 j0.4762 -0.3704 j0.3102 -0.3577 j0.3073
Frequency (Hz) 0.08 0.08 0.53 0.43 1.20 1.16 0.60 0.10 0.09 0.07 0.08 0.05 0.05
Damping ratio % 99.99 99.99 97.97 -98.63 10.05 10.23 3.22 99.72 99.76 89.08 73.87 76.67 75.85
Table 3.3 Damping Inter-Area Mode Comparison for Test System
Mode No Operating Conditions Frequency (Hz) Damping ratio % 13,14 Without SVC 0.49 2.98 13,14 With SVC 0.6 3.22 13,14 SVC with POD 0.5913 20
54

Response to severe disturbance
3
2.5
B s 9 v ltag m g itu e in p u o e a n d u
1.5
0.5
-0.5
SVC without POD With no SVC SVC with POD
-1 0
10
Time in seconds
Figure.3.5 Bus 9 Voltage Response

Response of system w ith severe disturbance
200
150
C r e t inlin 1 to9 ur n e 0
100
50
0 SVC with out POD -50 No SVC SVC with POD
-100
4 5 6 Time in seconds
10
Figure.3.6. Current response between buses 9 to bus10

Response w ith severe disturbance
5 4 3
S e dr s o s o G pe ep ne f 1
2 1 0 -1 -2 -3 -4 -5 SVC without POD No SVC SVC with POD controller
Tim e in seconds
10
12
14
16
18
20
Fig.3.7 Speed response of G1

Response w ith severe dsturbance
15
55
10
A ngular response of G (deg) 1
-5
-10
SVC without POD controller No SVC SVC with POD controller
-15
-20
10
Figure.3.8 Angular response of G1

Response of system w ith severe disturbance
1500
1000
G A t ep w r inM 1 civ o e W
500
-500 SVC without POD No SVC SVC with POD
-1000
-1500
10
Figure.3.9 active power response of G1

20
Response of system with severe disturbance SVC with PI Controller SVC with proposed POD SVC only
15
Magnitude of line currents
10
-5 0
10
Time(s)
Figure.3.10 Line current response from bus10-9
56
3.6 SIMULATION RESULTS The effectiveness of the proposed method of POD designed was tested on two- area four -machine systems. To verify the behavior of the proposed controller, a three-phase short circuit is applied to one of the transmission lines connecting bus 8 and bus 9. The fault occurs at 1.0 sec. and is cleared 74 ms later by opening the breakers at both ends. The line is then reconnected successfully to the system at 5.0 sec. The response of the system to such a disturbance is illustrated in Figure 3.7 to Figure 3.9. The results show that the SVC with POD stabilizes Figure 3.5 to Figure 3.9 and the oscillations of the system are well damped with POD controller as compared with SVC only or no SVC. A Hankel singular value (HSV) analysis was carried out to find the suitable feedback signals for the SVC controllers. Line109 line current which is a local signal has been taken as the feedback signal for the TCSC supplementary controller from HSV analysis. From the table 3a the damping increase from 2.98% to 20% as desired, this is more accurate then turning the stabilizer using frequency or root locus techniques. A PI controller was also used to compare the effectiveness of the method and the result is shown in Figure 3.10. It can be seen that the proposed controller provides a better and faster damping to the system oscillations as compared to the PI controller. 3.6 CONCLUSION This work has designed a power oscillation damping controller (POD) to control of power system oscillations with SVC device based on Residue, which is the product of controllability and observability of the linear model of the power system. This Residue method also provides valuable information on how to design power system damping controllers. Although eigenvalue based methods are very powerful, the complexity of the power system stability problem requires the complementary use of other methods such as non-linear time domain simulation. All the simulations were done with PSAT toolbox in Matlab environment.
57
REFERENCES
[1] B. Kalyan Kumar, S.N. Singh and S.C. SrivastavaPlacement of SVC controllers using modal controllability indices to damp outpower system oscillations IET Gener. Transmission. Distrib., Vol. 1, No. 2, page 209-217, March 2007 [2] R. J. Piwko, Ed., Applications of Static Var Systems for System Dynamic Performance, IEEE Publication 87TH018755-PWR, [3] N.G. Hingorani, Flexible ac transmission, lEEE Spectrum, April 1993, pp. 40-45. [4] M. A. Pai and Alex Stankovic Robust control in power system tex book. USA, Springer Science+Business Media, Inc. 2005 [5] Yong Chan, Design of HVDC and SVC Coordinate Damping Controller Based on Wide Area Signal International Journal of Emerging Electric Power Systems Vol .7, No. 4 2006 [6] K.R. Padiyar Power system dynamics stability and Control Anshan limited UK 2004 [7] R. Sadikovic, Damping controller design for power system oscillations, Internal report, ETH Zrich, December 2004. [8] E. Acha, C. Fuerte-Esquivel, H. Ambriz-Perez and C. AngelesCamacho. FACTS Modelling and Simulation in Power Networks, John Wiley & Sons LTD, England, 2004. [9] Bikash Pal , Balarko Chaudhuri Robust Control in power System Springer Science+Business Media, Inc.USA 2005 [10] kamoto, H., Kurita, A., and Sekine, Y.: A method for identification of effective locations of variable impedance apparatus on enhancement of steady state stability in large scale power systems, IEEE Trans. Power Syst., 1995, 10, (3), pp. 14011407 [11] M.E. Aboul-Ela, A.A. Salam, J.D. McCalley and A.A. Fouad, Damping Controller Design for Power System Oscillations Using Global Signals, IEEE Transactions on Power Systems, Vol. 11, No.2, May 1996, 767-773.
58
[12] Kundur P. Power system stability and control. New York, USA: McGraw- Hill; 1994.
Optimal Location of FACTS Devices for Damping Oscillations
59
4
OPTIMAL LOCATION OF FACTS DEVICES FOR DAMPING OSCILLATIONS
M.W. Mustafa N. Magaji F. Taha
4.1 INTRODUCTION Damping of electromechanical oscillations between interconnected synchronous generators is necessary for a secure system operation. While local oscillations involving one or more generators swinging against the rest of the system are mainly influenced by a restricted number of local system parameter- the behavior of low frequency inter-area oscillations is generally determined by global parameters of larger parts of the power system. Power System Stabilizers (PSSs) applied on selected generators can effectively damp local oscillation modes; their limited influence on inter-area modes, however, lead to the fact that they may not be considered as the only solution to damp interarea oscillations. Flexible AC Transmission Systems are being increasingly used to better utilize the capacity of existing transmission systems. Flexible AC Transmission System (FACTS) is a technology based solution to help the utility industry deal with changes in the power delivery business. A major thrust of FACTS technology is the development of power electric based systems that provide dynamic control of the power transfer parameters transmission voltage, line impedance and phase angle [1]. The static var compensator (SVC), which is a shunt device, provides dynamically variable shunt impedance to regulate the
60
voltage at a bus where it is connected. The thyristor-controlled series compensator (TCSC), which is a Series FACTS controller, is used to change the effective reactance of the line dynamically, thereby controlling the real power flow in the lines. The unified power flow controller (UPFC) is used to control the real and reactive powers simultaneously in a line by injecting a series voltage across the line. It also injects a shunt reactive current at one of the line-end buses to maintain the bus voltage at a specified value [2, 3]. FACTS devices are provided with supplementary controllers for damping out small signal oscillations [4-8], and the effectiveness of damping the oscillations depends on the location of FACTS controllers. Several methods [24] have been proposed for the placement of FACTS controllers. Previous studies [910] considered only static criterion like improving power transfer, available transfer capability (ATC), loss minimizations and did not consider any dynamic criteria for the placement of the FACTS controllers. A sensitivity based approach has been developed in some researches for determining the optimal placement of FACTS devices in electricity market having pool and contractual dispatches [11-13]. [11] Shows that few locations of FACTS devices can be decided based on sensitivity factors and then optimal dispatch problem must solved for finding suitable placement of UPFC in a congested system. [15] Implements Genetic Algorithm to finding the best location for TCSC to maximizing total transfer capacity. Tuning of FACTS parameters has been proposed for improving damping of weakly damped inter-area modes. [16-17] proposed a method for selecting suitable feedback signal to FACTS controllers for improving the damping. Sadikovic and Korba [18] used residue, called location index for effective damping, to find suitable location for damping inter-area mode of oscillations, also in the work of [14] and [19], the method used only UPFC placement based on only single operating condition. No consideration of other FACTS devices.
61
In this work, a residue technique has been proposed for optimal placement of FACTS controllers similar with the work of [18] with the following modifications: 1 More than one test system is taken into consideration in this work starting from small case study to a little bigger case study. 2 The number of FACTS devices used in this work cover the classification of FACTS devices in terms of series and parallel and in terms of fabrications of the devices. 3 In this work one of the primary functions of FACTS devices that are available transfer capability is also taken into consideration by finding the placement of FACTS controllers through a static criterion based on loss sensitivity indices. This location shows different location from the location due to damping oscillations. The proposed method has been tested on two area 11 bus systems and New England 39-bus system. The effectiveness of the proposed method has been demonstrated through the time domain simulation
4.2 RESIDUE FACTOR Let start from the mathematical model a dynamic system expressed in terms of a system of non-linear differential equations: x = F ( x, t ) (1) If this system of non-linear differential equations is linearized around an operating point of interest x=x0, it results in: x = Ax(t ) (2) Assume that an input u(t) and an output y(t) of the linear dynamic system equation (2) have defined
x(t) = Ax(t) + Bu(t) y(t) = Cx(t)
(3)
Considering equation (3) with single input and single output (SISO) and assuming D = 0, the open loop transfer function of the system can be obtained by
G( s ) = y( s ) u( s )
1
(4)
= C( sI A ) B
62
The transfer function G(s) can be expanded in partial fractions of the Laplace transform of y in terms of C and B matrices and the right and left eigenvectors as N C B G( s ) = i i i =1 ( s i ) N (5) Rijk = i =1 ( s i ) Where Rijk is the residue factor associated with ith mode, jth output and kith input. Rijk can be expressed as: (6) Where vi, and i denote the right and left eigenvectors, respectively associated with the ith eigenvalue. This can be expressed in terms of mode controllability arid observability. The controllability of mode i from the k-th input is given by: (7) The measure of mode observability of mode i from output j is given by: Obsvij = C j vi (8) It is clear that
Rijk = C j vi wi Bk = obsvij cont ik
CI ik = wi Bk
Rijk = C j vi wi Bk
(9)
Each term in the denominator, Rijk, of the summation is a scalar called residue. The residue Rijk of a particular mode i give the measure of that modes sensitivity to a feedback between the output y and the input u; it is the product of the modes observability and controllability.
63
Figure 4.1 Two-area four machine test system
Figure 4.1 shows a system G(s) equipped with a feedback control H(s). When applying the feedback control, eigenvalues of the initial system G(s) are changed. It can be proven [20], that when the feedback control is applied, the shift of an eigenvalues can be calculated by i = Ri H( i ) (10) It can be observed from (10) that the shift of the eigenvalue caused by the controller is proportional to the magnitude of the corresponding residue. For a certain mode, the same type of feedback controls H(s), regardless of its structure and parameters can be tested at different locations. For the mode of the interest, residues at all locations have to be calculated. The largest residue then indicates the most effective location to apply the feedback control [21]. The FACTS controllers have been placed on the basis of the maximum value of the residue corresponding to the most critical mode. In this study, only inter-area modes of oscillations have been considered. An SVC has been placed at a bus having maximum value of the residue. A TCSC has been placed in the line having maximum value of the residue .For the placement of a UPFC, first the placement of the series converter has been decided in a line having maximum residue, according to its series voltage injection, because controlling the real power in the line has more effect on the inter-area mode damping.
64
4.3 4.3.1
CRITERIA FOR OPTIMAL FACTS CONTROLLERS Dynamic criterion based on residue
PLACEMENT
OF
The FACTS controllers have been placed on the basis of the maximum value of the residue corresponding to the most critical mode. In this study, only inter-area modes of oscillations have been considered. A TCSC has been placed in the line having maximum value of the residue while SVC has been placed at a bus having maximum value of the residue. For the placement of a UPFC, first the placement of the series converter has been decided in a line having maximum residue, according to its series voltage injection, because controlling the real power in the line has more effect on the interarea mode damping. Thereafter, the bus at which the shunt converter of UPFC was to be connected (sending end or receiving end of the line) was decided as the bus having maximum residue
4.3.2 Static criterion based on loss sensitivity indices
The FACTS controllers are used for increasing power transfer capability and minimize transmission losses, apart from damping the system oscillations. This will require use of certain static criterion. In general, the optimal location of the FACTS controllers obtained from a dynamic criterion is not same as that with a static criterion. Few authors[21] have tried to define a hybrid participation factor to optimally place the FACTS controllers so as to determine a compromising optimal location to achieve both static as well as dynamic performance enhancement. However, such methods use relative weights to the static and dynamic participation factors to
65
In this work, the static criterion used for optimal location of FACTS controllers is based on a loss sensitivity index as defined in the work of [2, 3] and given as follows. The loss sensitivity index k for SVC placed at bus i, is defined as the sensitivity of the total transmission losses with respect to the reactive power injected at a bus i
k = 2
j =1 N
Rij ViV j
(cos (
j ) Q j + sin ( i j ) Pj
(11)
Where Rij the real part of ijth element of the Z bus, N is number of buses, Pj and Qj the real and reactive power injections at j th bus. The loss sensitivity index k , for a TCSC placed in a line l, between buses i and j, is defined as the sensitivity of the total transmission losses with respect to the line net series reactance
k =
(R
2 Rk X k
2 k
+ X k2
(V + V
i
2ViV j cos ( i j )
(12)
Where Rk and Xk are the line resistance and reactance, respectively. The loss sensitivity index k , for UPFC placed in the line l between buses i and j, is defined as the sensitivity of the total transmission losses with respect to complex part of the series injected voltage in the line l
k = Vi V j cos ( i j )
)R X X +
k 2 k
2 k
R V j sin ( i j ) 2 k 2 Rk + X k
(13)
Optimal location of the SVC, utilizing static criterion, was considered at a bus having most negative value of k, whereas TCSC and UPFC were placed in a line having maximum absolute value of k and k, respectively.
4.4 RESULTS AND DISCUSSION
The effectiveness of the proposed method was tested on two areafour machine eleven bus system and New England 10 machine, 39-
66
bus system. The results for the two systems are presented in what follows.
4.4.1 Two areas four machine In this study, a two area interconnected four machine power system shown in Figure 4.1 is considered. The system consists of four machines arranged in two areas inter-connected by a weak tie line[22]. The system is operating with area 1 exporting 400 MW to area 2. The equivalent circuit of the system is shown in Figure 4.2.
Figure 4.2 Two-area four machine test system
Network and generator data can be found in [22]. We have assumed that the generators are equipped with static excitation systems. Modal analysis shows that, in the base case, the critical mode is an interarea mode having eigenvalues, -0.07934 +j6.223 and damping ratio 0.0127, which is relatively low. Analysis of phase angle of right eigenvectors shows that in this interarea mode, generator G2 and G2 oscillates against a cluster of generators G3, G4. The proposed Residue factor, corresponding to the weakly damped inter-area mode, computed for SVC, TCSC and UPFC are given in Table 4.1 to 4.3.
67
Table 4.1 Residue Factor of Inter: Area Oscillating Mode for UPFC (11Bus System)
Line 7--8 6--7 5--6 9--8 10--11 10--9
Base 0.4202 0.9406 0.5026 0.3533 0.5718 1.0000
Operating Conditions Line 7-8 Line 9-8 Outage Outage 0.0001 0.4944 1.0000 1.0000 0.4501 0.4531 0.6606 0.0000 0.4914 0.5042 0.8924 0.9222
Average 0.3049 0.9802 0.4686 0.3380 0.5225 0.9382
Table 4.2 Residue Factor of InterArea Oscillating Mode For TCSC (11Bus System)
Operating Conditions Line Base Line 7-8 Outage Line 9-8 Outage 7--8 0.21 0.618321 0.259 6--7 0.51 1 1 5--6 0.25 0.267176 0.269 9--8 0.39 0.251908 0.641 10--11 1 0.21374 0.223 10--9 0.7625 0.763359 0.8 Average 0.3624 0.8367 0.2621 0.4276 0.4789 0.7753
68
Table 4.3 Residue Factor of Interarea oscillating mode for SVC (11 bus system)
Bus 5 6 7 8
Operating Conditions base Line 7-8 outage 0.015131 0.009149 0.000004 0.631615 0.602169 1.000000 0.733657 0.605421 0.369030 0.011844
Line 9-8 outage 0.459467 0.000018 0.000017 1.000000 0.000007 0.000004 0.492076
Average 0.1612 0.2105 0.5341 0.9112 0.3331 0.1230 0.1750
1.000000 9 0.393927 10 0.000004 11 0.021197
Table 4.4 Facts Placement using normalized Loss sensitivity index 11bus system
UPFC Line 7-8 6-7 5-6 9-8 10-11 Loss function 0.4102 0.2871 0.2518 0.2122 1.0000 Line 7-8 6-7 5-6 9-8 10-11
TCSC 0.4029 0.0165 0.000026 0.7626 0.0077 6 7 8 9 10
SVC -0.0089 0.0000 -0.0833 0.0000 -0.0052
Loss function Bus Loss function
These indices are expressed in the normalized form. All buses/lines, having significant value of the residue factor, are listed in these tables except the generator buses and transformer branches.
69
Table 4.5 Eigenvalue of critical inter-area mode without and with FACTS in the base case for11 bus system
Free fault No FACTS UPFC SVC TCSC
Eigenvalue -0.089j3.5618 -0.094j3.6762 -0.095 j3.677 -0.178j1.682
Damping Ratio (p.u) 0.025 0.0255 0.0258 0.1055
Frequency (Hz) 0.567 0.585 0.585 0.268
It is observed from Table 4.1 to 4.2 that the line between buses 7 8, the line between buses 10-11 and bus 8 are having the maximum residue factor values for TCSC, UPFC and SVC placement, respectively. Optimal location of these FACTS controllers, based on static criterion, was also found using loss sensitivity indices. These indices computed for SVC, TCSC and UPFC placement in the normalized form are given in Table 4.4. A Hankel singular value (HSV) analysis was carried out to find the suitable feedback signals for the FACTS controllers [25]. Line78 real power flow in which TCSC is placed, a local signal, has been taken as the feedback signal for the TCSC supplementary controller. Current magnitude of Line 78 from bus 7 is taken as the feedback signal for the UPFC supplementary controller. While for SVC the bus voltage of bus 10 is taken as feedback signal for the supplementary controller.
4.4.2 Performance Evaluation in 11 bus system
A nonlinear time-domain simulation has been carried out for the 11 bus study system by applying a step input of 1.0 per unit to the generator G1 exciter reference voltage, in order to evaluate the impacts of SVC, TCSC and UPFC on the stable inter-area mode. The rotor states of generator G1 are the most connected states in the inter-area mode according to eigenvalue analysis.
70
Response of FACT S controllers for 11 bus sysem

0.2
0.15
No FACTS
With TCSC
With SVC
With UPFC
0.1
S e d v tio o G p d e ia n f 1
0.05
-0.05
-0.1
-0.15
-0.2
10
12
14
16
18
20
Time(s)
Figure 4.3.
Speed deviation of G1 for 11 bus system
Figure 4.3 shows the speed deviation of generator G1 without FACTS controllers and with SVC, TCSC and UPFC, placed one at a time in the system. It can be observed from Table 4.5 and Figure 4.3 that the inter-area mode, without any FACT device, has the eigenvalues very close to the imaginary axis. Hence, it exhibits an oscillatory response for a step input and oscillates even after 20s. It can be seen that oscillations are damped out after SVC placement at bus 8 in about 18s. With TCSC placed in Line 6-7, the oscillations are damped out in about 14s; it can be observed from this figure that the oscillations in both the cases extend even after 10s, since the frequency of the inter-area mode of oscillation is about 0.57 Hz. It is also observed from Figure 4.1 that the oscillations are damped in about 2s with UPFC placement, which is quite less when compared with the TCSC and SVC placement. The impact on damping of the critical inter-area mode with and without FACTS controllers is given in table 4.5.
71
4.4.3 New England 39-bus system
This system [20] consists of 10 generators, with generator 2 taken as reference generator. The single-line diagram of the New England system is given in Figure 4.4. Modal analysis shows that, in the base case, the critical mode is an interarea mode having eigenvalues, -0.07934 +j6.223 and damping ratio 0.0127, which is relatively low. Analysis of phase angle of right eigenvectors shows that in this interarea mode, generator G31 oscillates against a cluster of generators G32, G33 and G34.
Figure 4.4: The New England 39- bus system
72
The proposed residue, corresponding to the weakly damped interarea mode, computed for SVC, TCSC and UPFC are given in Table 4f.These indices are expressed in the normalized form. Only few buses/lines, having significant value of the residue, are listed in this table and these do not include generator buses and transformer branches. It is observed from Table 4.7 that the line between buses 1621, the line between buses 34 and bus 16 are having the maximum residue values for UPFC, TCSC and SVC placement, respectively. Optimal location of these FACTS controllers, based on static criterion, was also found using loss sensitivity indices. These indices computed for SVC, TCSC and UPFC placement in the normalized form are given in Table 4.8.
Table 4.7 Residue factor of FACTS devices in the base case for 39-bus system
Line 9-39 17-27 15-16 16-21 14-15 21-22 16-17 2-3 3-4 2-25 7-8 25-26 26-28 28-29 17-18
TCSC Normalize d residue 0.34 0.59 0.28 1.00 0.26 0.58 0.41 0.53 0.13 0.68 0.28 0.32 0.22 0.22 0.45
UPFC Normalized residue 0.240 0.129 0.158 0.138 0.394 0.188 0.097 0.190 1.000 0.305 0.192 0.075 0.079 0.253 0.262
Bus 28 24 27 21 18 17 16 15 14 9 8 7 6 5 4
SVC Normalize d residue 0.561 0.363 0.913 0.080 0.191 0.494 1.000 0.053 0.215 0.660 0.863 0.624 0.220 0.316 0.268
73
Table 4.8 Facts Placement using normalized Loss sensitivity index 39bus system
SVC Bus Index 16 -0.11 19 -0.13 25 -0.13 28 -0.18 24 -0.2 17 -0.23 22 -0.25 5 -0.25 29 -0.28 26 -0.3 3 -0.44 2 -0.56 9 -1
TCSC UPFC Line Index Line Index 11-12 0.01 28 -29 0.089 13-12 0.01 18-3 0.119 13 -14 0.013 2-10 0.119 21- 22 0.014 16 19 0.124 16 -19 0.021 21 22 0.174 3-2 0.023 3-2 0.212 26- 29 0.06 12 11 0.309 1-9 0.092 12 13 0.322 4-3 0.109 1-9 0.435 25- 26 0.137 1-10 0.448 1-10 0.192 4-3 0.688 23- 24 0.308 23 24 0.768 9-8 1 9-8 1
4.4.3 Performance Evaluation in 39 bus system
A nonlinear time-domain simulation has been carried out for the 39 bus study system by applying a step input of 1.0 per unit to the generator G31 exciter reference voltage, in order to evaluate the impacts of SVC, TCSC and UPFC on the stable inter-area mod. The rotor states of generator G31 are the most associated states in the inter-area mode according to eigenvalue analysis. Figure.4.5 shows the speed deviation of generator G31 without FACTS controllers and with SVC, TCSC and UPFC, placed one at a time in the system.
74
Step Response of 39 bus sysem

1.5
With UPFC
1
With SVC
With TCSC
Without FACTS
Speed deviation of G2
0.5
-0.5
-1
10
15
20
25
Time(s)
Figure 4.5
Speed deviation of G31 for 39 bus system
Table 4.9 Eigenvalue Of Critical Inter-Area Mode Without And With Facts In The Base Case For 39 Bus System
Bus 39 NO.FACTS SVC TCSC UPFC
Eigenvalue 0.14604j5.4322 0.14932j4.9395 0.14261j4.4714 0.27622j5.4622
Damping Frequency (Hz) Ratio 0.027 0.03 0.032 0.051 0.864 0.786 0.711 0.869
It can be observed from Table 4.9 and Figure. 4.5 that the interarea mode, without any controller, has the eigenvalues very close to the imaginary axis. Hence, it exhibits an oscillatory response for
75
a step input and oscillates even after 25 s. It can be seen that oscillations are damped out after SVC placement at bus 16 in about 20s, even though the initial overshoot has increased. With TCSC placed in Line 2122, the oscillations are damped out in about 10s; it can be observed from this figure that the oscillations in both the cases extend even after 10s, since the frequency of the inter-area mode of oscillation is about 0.7 Hz. It is also observed from Figure.4.5 that the oscillations are damped in about 5 s with UPFC placement, which is quite less when compared with the TCSC and SVC placement. The impact on damping of the critical inter-area mode with and without FACTS controllers is given in Table 4.9.
4.5 CONCLUSION Residue factor for the placement of FACTS controllers have been found corresponding to a critical inter-area mode in case of the 11bus and 39-bus systems. These factors have been computed for the base case only. The eigenvalue analysis and the results of the step response show that all the three FACTS controllers provide significant enhancement in the damping of oscillations REFERENCES
[1]
[2]
[3]
H. A. A. Kazemi, "FACTS Devices in Deregulated Electric Power Systems:Review," in IEEE International Conference on Electric Utility Deregulation, Restructuring and Power Technologies, Hong Kong, April 2004 M. N. A. R Messina, "An Efficient Placement Algorithm Of Multiple Controllers For Damping Power System Oscillations," IEEE Transactions on Power Systems, pp.. 1280-1285., 1999. S. N. S. and. S. C. S. B. Kalyan Kumar, "Placement of FACTS controllers using modal controllability indices to damp out power system oscillations," IET Gener. Transm. Distrib., vol. 1, pp. 209217., 2007.
76
[4]
[5]
[6] [7] [8]
[9] [10] [11]
[12]
X. R. Chen, Pahalawaththa, N.C., Annakkage, U.D., and Kumble, C.S., "Output feedback TCSC controllers to improve damping of meshed multi-machine power systems," IEE Proc. Gener. Transm. Distrib, vol. 144, pp. 243248, 1997. H. Okamoto, Kurita, A., and Sekine, Y., "A method for identification of effective locations of variable impedance apparatus on enhancement of steady state stability in large scale power systems," IEEE Trans. Power Syst., vol. 10, pp. 14011407, 1995. J. J. S.-G. and. J. H. C. E.V. Larsen, "Concepts for Design of FACTS Controllers to Damp Power Swings," IEEE transactions on Power Systems, vol. 10, pp. 948-955, 1995. K. M. Son, and Park, J.K, "On the robust LQG control of TCSC for damping power system oscillations," IEEE Trans. Power Syst., vol. 15, pp. 13061312, 2000. M. G. M. Noroozian, G. Andersson, J. Gronquist and I. Hiskens, "A Robust Control Strategy for Shunt and Series Reactive Compensators to Damp Electromechanical Oscillations," IEEE transactions on Power Delivery, vol. 16, pp. 812-817. , 2001. R.S. Fang and A.K. David, "Optimal dispatch under transmission contracts," IEEE Transaction. Power System, vol. 14, pp. 732-737, May 1999. S. N. Singh, and David, A.K, "A new approach for placement of FACTS devices in open power market," IEEE Power Eng. Rev, vol. 21, pp. 5860, 2001. S. N. Singh, and David, A.K, "Placement of FACTS devices in open power market" in IEEE Proceedings of the 5th International Conference on Advances in Power System Control, Operation and Management, APSCOM 2000,, Hong Kong,, October-November 2000., pp. 173 -177. S. N. Singh, and David, A.K,, "Congestion management by optimizing FACTS device location," in International Conference on Electric Utility Deregulation and
77
[13]
[14]
[15]
[16] [17]
[18]
[19]
[20]
[21]
Restructuring and Power Technologies ; 2000 Proceedings., DRPT, 4-7 April 2000, pp. 23 -28. G. B. Wang Feng; Shrestha, "Allocation of TCSC devices to optimize total transmission capacity in a competitive power market," Power Engineering Society Winter Meeting 2001. IEEE, vol. 2, pp. 587 -593, 28 Jan.-1 Feb. 2001. B. Chaudhuri, and Pal, B.C., "Robust damping of multi swing modes employing global stabilizing signals with a TCSC," IEEE Trans. Power System., vol. 19, pp. 499505, 2004. M. M. Farsangi, Song, Y.H., and Lee, Y, "Choice of FACTS device control input for damping inter-area oscillations," IEEE Trans. Power System, vol. 19, pp. 11351143, 2004. P. K. R. Sadikovic, G. Andersson., "Application of FACTS Devices for Damping of Power System Oscillations," Proc. of IEEE Power Tech St. Petersburg, Russia,, 2005. K. S. S. Verma, S.N.; Gupta, H.O., "FACTS devices location for enhancement of total transfer capability," Power Engineering Society Winter Meeting IEEE, vol. 2, pp. 522 -527, 28 Jan.-1 Feb 2001. P. K. R. Sadikovic, G. Andersson, IEEE, "Self-tuning Controller for Damping of Power System Oscillations with FACTS Device ETH Zurich, Switzerland. sadikovic@eeh.ee.ethz.," 2006. A. A. S. M.E. Aboul-Ela, J.D. McCalley and A.A. Fouad, "Damping Controller Design for Power System Oscillations Using Global Signals," IEEE Transactions on Power Systems, vol. 11 pp. 767-773, May 1996. R. Sadikovi C, "Use Of Facts Devices for Power Flow Control and Damping of Oscillations in Power Systems." PhD Zurich: Swiss Federal Institute of Technology 2006, p. 139. C. P. Gupta, Srivastava, S.C., and Verma, R.K. , "Static and dynamic criterion of SVC placement for voltage
78
[22]
stability enhancement," in Int. Conf. Power Systems ICPS Katmandu, Nepal, 2004. P. Kundur, Power system stability and control. New York,USA: McGraw- Hill, 1994.
Eigenvalues Based Method for Analysis of Power System With TCSC Device for Damping Oscillations
79
5
EIGENVALUES BASED METHOD FOR ANALYSIS OF POWER SYSTEM WITH TCSC DEVICE FOR DAMPING OSCILLATIONS
M.W. Mustafa N. Magaji Z. Muda
5.1 INTRODUCTION The concept of flexible ac transmission systems is made possible by the application of high power electronic devices for power flow and voltage control [l]. In addition a number of TCSC devices have already been installed to aid power system dynamics which help to mitigation a low frequency oscillations often arise between areas in a large interconnected power network [2]. Eigenvalue sensitivities are one important outcome of the modal analysis and control of oscillatory behavior and dynamic stability in power systems. The pioneering work [3] considers the local oscillation of a single machine by means of a transfer function model. The usually complex pattern of oscillations in a large power system can be studied through linear, time invariant, statespace models based on the perturbations of the system state variables from their nominal values at a specific operating point Power system oscillations occur due to the lack of damping torque at the generators rotors. The oscillation of the generators rotors cause the oscillation of other power system variables (bus voltage, bus frequency, transmission lines active and reactive powers, etc.). Power system oscillations are usually in the range between 0.1 and 2 Hz depending on the number of generators involved in [4]. Local oscillations lie in the upper part of that range and consist of the
80
oscillation of a single generator or a group of generators against the rest of the system. In contrast, inter-area oscillations are in the lower part of the frequency range and comprise the oscillations among groups of generators. In addition, power system oscillations exhibit low damping compared to oscillations found in other dynamic systems: an oscillation of 10% damping is commonly accepted as well damped. To improve the damping of oscillations in power systems, supplementary control laws can be applied to existing devices. These supplementary actions are referred to as power oscillation damping (POD) control This work reviews the basic concepts of eigenvalue analysis of linear systems. The physical meaning of eigenvalues, eigenvectors, participation factors, residues and controllability and observability indices will be introduced and illustrated in small scale power systems. This technique has been successfully used in location and tuning of power system stabilizers [5] and FACTS devices. The application of sensitivity measures to the design of power system damping (POD) controllers has been applied to TCSC. The design method utilizes the residue approach; this presented approach solves the optimal sitting of the TCSC device, selection of the proper feedback signals and the controller design problem [6].
LINEAR SYSTEM ANALYSIS TOOLS IN POWER SYSTEMS Low frequency electromechanical oscillations range from less than 1 Hz to 3 Hz other than those with sub-synchronous resonance (SSR) [6,15]. Multi-machine power system dynamic behavior in this frequency range is usually expressed as a set of non-linear differential and algebraic (DAE) equations. The algebraic equations result from the network power balance and generator stator current equations. The initial operating state of the algebraic variables such as bus voltages and angles are obtained through a standard power flow solution. The initial values of the dynamic variables are obtained by solving the differential equations 5.2
81
5.2.1 Eigenvalues, Eigenvectors and Modes . Let us start from the mathematical model a dynamic system expressed in terms of a system of non-linear differential equations:
x = F ( x, t ) (1) If this system of non-linear differential equations is Linearized around an operating point of interest x=x0, it results in: x = A x (t ) (2) A meaningful solution method of equation (2) is based on the eigenvalues and eigenvectors of the state matrix A. An eigenvalue i of the state matrix A and the associated right vi and left wi eigenvectors are defined according to: Avi = i vi In a matrix with all distinct eigenvalues (not a necessity but it is easier to understand when it is so), all the right eigenvectors and eigenvalues can be expressed as a compact matrix expression AV = VA (3) where, V = ( v1 ,v2 ..... vn 1 vn , ) (4) = diag ( 1 2 ...n-1 ,n ) Pre-multiplying both sides of (3) by V-1 gives V 1 AV = (5) A similar expression holds for the left eigenvectors W such that WA = W (6) Where t t t W = [ 1t ,2 ..... n 1 n ]t (7) -l Post-multiplying both sides of (6) by W , gives WAW 1 = (8) The transformed physical state variables (x) can be put into modal variables equation (2) with the help of eigenvector matrices V and W
x = Vz z = Wx
(9)
82
In power system literature, the right eigenvector matrix v is known as the mode shape matrix, that is, eigenvector vi is known as the ith mode shape, corresponding to eigenvalue i. The mode shape provides important information on the participation of an individual machine or a group of machines in one particular mode. solution of equation (2) can be expressed in terms of the eigenvalues and eigenvectors of the state matrix as:
x(t) = Ve t Wx(0) = vi ei t [w iT x(0)]
i =1 N
(10)
The analysis of equation (10) allows to draw the following conclusions: i. The system response is the combination of the system response to each of the N modes. ii. The eigenvalues determine the system stability. A real positive (negative) eigenvalue determines exponentially increasing (decreasing) behavior. A complex eigenvalue of positive (negative) real part results in a increasing (decreasing) oscillatory behavior. iii. The components of the right eigenvector vi measure the relative activity of each variable in the ith mode. iv. The components of the left eigenvector wi weight the initial conditions in the i-th mode
5.2.2 Participation factors It is natural to suggest that the significant state variables influencing a particular mode are those having large entries corresponding to the right eigenvector of i. The participation factor of the j-th variable in the k-th mode is defined as the product of the j-th's components of the right vjk and left wki eigenvectors corresponding to the k-th mode [7] Pjk =W jkVkj (11) The product W jkVkj is a dimensionless measure which is called participation factor. In other words, they are independent on the units of the state variables. In addition, both the sum of the participation factors of all variables in a mode and the sum of the participation of all modes in a variable are equal to one. Other
83
interesting measure is the subsystem participation. The subsystem participation is the magnitude of the sum of the participation factors of the variables that describe a subsystem in a mode.
5.2.3 Modal controllability and observability factors The effectiveness of control in power system can be indicated through controllability and observability indices. This is important as control cost is influenced to a great deal by the controllability and observability of the plant. These issues are addressed through modal controllability and observability 5.2.4 Controllability index Assume that an input u(t) and an output y(t) of the linear dynamic system (2) have defined
x(t) = Ax(t) + Bu(t) y(t) = Cx(t)
(12)
The application of a linear transformation defined by the eigenvectors of the state matrix to the system as described by (12) results in: equation (13): Let v and w be the right and left eigen vector matrices of A, respectively. If eigenvalues of A are distinct, then wTv = I, where wT is conjugate transpose of w and I is the identity matrix. Substituting x =wz in (12), we obtain
(13)

z( t ) = k z k ( t ) + w T Bi vi ( t ) k
i =1 m
(14)
Where wk is the left eigenvector corresponding to kth mode and Bi is the ith column vector of matrix B. From (14), one can find the controllability of kth eigen mode with respect to the ith input. The controllability index (CI) of an ith input to the kth mode [8] is defined as CIi = w T Bi (15) k
84
T The input i, for which the value of wk Bi is maximum, is considered the suitable parameter to be controlled for affecting the kth eigen mode to maximum extent.
Observability index The observability index (cvi) of an ith input to the kth mode is defined as OIi = Ci wk (16) The study of equations (15) and (16) leads to the following conclusions: CIi Measures the controllability of the mode associated to i. the variable x i ( t ) from the input u( t ) .In other words, if the mode i can be controlled from the input u( t ) ii. OIi Measures the observability of the mode associated to the variable x i ( t ) form the output y( t ) . In other words, if the mode i can be observed from the variable y( t ) Therefore, a mode can be controlled if only if it is controllable from the input u( t ) and observable from the output y( t )
5.2.5
5.3 RESIDUES Considering (12) with single input and single output (SISO) and assuming D = 0, the open loop transfer function of the system can be obtained by
G( s ) = y( s ) u( s )
1
(17)
= C( sI A ) B
G( s ) = =
N
C i B i i =1 ( s i )
N
Ri ( s i ) i =1
(18)
85
Each term in the denominator, Ri, of the summation is a scalar called residue. The residue Ri of a particular mode i gives the measure of that modes sensitivity to a feedback between the output y and the input u; it is the product of the modes observability and controllability. Figure 5.1 shows a system G(s) equipped with a feedback control H(s). When applying the feedback control, eigenvalues of the initial system G(s) are changed. It can be proven, that when the feedback control is applied, the shift of an eigenvalues can be calculated by i = Ri H( i ) (19) It can be observed from (19) that the shift of the eigenvalue caused by the controller is proportional to the magnitude of the corresponding residue. For a certain mode, the same type of feedback controls H(s), regardless of its structure and parameters can be tested at different locations. For the mode of the interest, residues at all locations have to be calculated. The largest residue then indicates the most effective location to apply the feedback control.
5.3. TCSC MODEL Thyristor-controlled series capacitor (TCSC) is a series FACTS device which allows rapid and continuous changes of the transmission line impedance It has great application potential in accurately regulating the power flow on a transmission line, damping inter-area power oscillations, mitigating sub synchronous resonance (SSR) and improving transient stability. A typical TCSC module consists of a fixed series capacitor (FC) in parallel with a thyristor controlled reactor (TCR) as shown in Figure 5.1. The TCR is formed by a reactor in series with a bi-directional thyristor valve that is fired with an angle ranging between 900 and 1800 with respect to the capacitor voltage [9]
86
vc
itcr
c
ltcr
TCR
Figure. 5.1 TCSC Model
The model to be adopted for any device in power systems analysis must be in accordance with the type of study involved and the tools used for simulation. Since this work is concerned with the application of the TCSC for stability improvement, the TCSC model used must rely in the assumptions that are typically adopted for transient stability analysis, i.e., voltages and currents are sinusoidal, balanced, and operate near fundamental frequency. In [9], a TCSC model suitable for voltage and angle stability applications and power flows studies is discussed. In that model, the equivalent impedance Xe of the device is represented as a function of the firing angle , based on the assumption of a sinusoidal steady-state controller current. The TCSC is modeled here as a variable capacitive reactance within the operating region defined by the limits imposed by the firing angle . Thus, Xemin Xe Xemax, with Xemax = Xe(min) and Xemin = Xe(180o) = XC, where XC is the reactance of the TCSC capacitor. (In this work, the controller is assumed to operate only in the capacitive region, i.e. min > r, where r corresponds to the resonant point, as the inductive region associated with 90o < < r induces high harmonics that cannot be properly modeled in stability studies [10]. The dynamic model characteristics of the TCSC are assumed to be modeled by a set of differential equations as follows (11) and model in Figure 5.2.
87
v POD
Kr
+ Pkm
+
KP + KI
max
1 Tr s+1
Pref
0
s
B( xc , )
min
Figure.5.2 small-signal dynamic model of TCSC
x 1 = ( 0 + K r v POD -x 1 )/Tr x 2 = K I (Pkm -Pref )
(20) (21) (22)
Where 0 = K P (Pkm -Pref ) + x2
The state variable x1=0, for firing angle model of TCSC. The PI controller is enabled only for the constant power flow operation mode [11]. According to D Jovic [12 ] the value of susceptance B is given as:
2 4 B( ) = k4 -2kx +1 coskx ( ) / xC kx coskx ( ) x
4 2 - coskx ( ) 2kx coskx ( ) + 2kx coskx ( ) 4 2 -kx sin 2 coskx ( ) + kx sin 2 coskx ( )
(23)
3 2 4kx cos2 sinkx ( ) -4kx cos sin coskx ( )
Where k X=XC/XL., The limits of the controller are given by the firing angle limits, which are fixed by design. 5.3.1 TCSC POD Controller Design Supplementary control action applied to TCSC devices to increase the system damping is called Power Oscillation Damping (POD). Since TCSC controllers are located in transmission systems, local
88
input signals are always preferred, usually the active or reactive power flow through TCSC device or TCSC terminal voltages. Figure 5.3 shows the considered closed-loop system where G(s) represents the power system including TCSC devices and H(s) TCSC POD controller In order to shift the real component of i to the left, SVC POD controller is employed. That movement can be achieved with a transfer function consisting of an amplification block, a wash-out block and mc stages of lead-lag blocks. We adapt the structure of
Yref ( s )
G(s)
Y ( s)
H ( s )TCSC
POD controller given in [13, 9] , i.e. the transfer function of the TCSC POD controller is
H ( s)TCSC
sTw 1+ sTlead 1 = K* * 1 + sTm 1+ sTw 1 + sTlag
mc
(24)
= KH1( s) Where K is a positive constant gain and H1 is the transfer function of the wash-out and lead-lag blocks. The washout time constant, Tw, is usually equal to 5-10 s. The lead lag parameters can be determined using the following equations:
comp = 1800 arg( Ri )
(25)
89
c =
Tlead Tlag
1-sin comp mc = 1+sin comp mc
(26)
Tlag =
i c
, Tlead = c Tlag
(27)
Where arg(Ri) denotes phase angle of the residue Ri, i is the frequency of the mode of oscillation in rad/=sec, mC is the number if compensation stages (usually mC = 2). The controller gain K is computed as a function of the desired eigenvalue location ides according to equation 26:
K=
i d Ri H1 (i )
(28)
5.4 EIGENVALUE ANALYSIS OF POWER SYSTEMS The concepts detailed in the previous section will be illustrated considering two small-scale power systems. The size of these systems allows the computation of all eigenvalues and
1 1 + sTm
KP
sTw 1 + sTw
1 + sTlead 1 + sTlag
1 + sTlead 1 + sTlag
eigenvectors of the state matrix without employing advance techniques due to small sizes of the system.
90
5.4.1 Analysis of single machine connected to an infinite bus with TCSC The case of a single generator connected to an infinite bus is considered first with and without TCSC. The generator model contains accurate representations of the synchronous machine, the excitation and the speed-governing systems. It has been assumed that the generator is equipped with a static excitation system [14]. A thyristor controlled series capacitor is connected between bus 2 and 3 as shown in Figure 5.5. The linear model of this system is described by 11 state variables. The synchronous machine, the TCSC and the exciter are described respectively by 6, 3 and 2 state variables. The eigenstructure of the state matrix contains 3 pairs of complex eigenvalues and 5 real eigenvalues which are detailed in tables 5.1 and 5.2 respectively. Eigenvalues accurately determine linear systems stability: this system is close to instability due to the presence of a poorly damped oscillatory mode. However, if the connections between eigenvalues and state variables are sought, participation factors have to be used. Table 5.3 details the participation of the generator subsystems: rotor dynamics, synchronous machine, exciter, and TCSC in all modes. Table 5.3 clearly indicates that the poorly damped oscillatory mode (eigenvalues 1 and 2) is associated to the rotor dynamics and that the other oscillatory mode (eigenvalues 3 and 4) describes the interaction between the synchronous machine and the exciter. The mode associated to the rotor dynamics is also known as electromechanical mode. Table 5.3 also shows that three exponential modes are associated to the machine (damper windings), other two to TCSC and the remaining mode to the exciter. The slower modes correspond to the TCSC dynamics whereas the fastest mode is associated to the exciter 5.4.2 Power system stabilizer Design of power system stabilizers or power oscillation damper (POD) in case of FACT can also be addressed using eigenvalue methods. Eigenvalue sensitivities with respect to the parameters of
91
the stabilizer provide a first order approximation of the eigenvalue movement in the complex plane when those parameters are varied. Precisely, the residue of the transfer function between the stabilizers output (reference of the excitation system, Vr) and the stabilizer input (speed, w, terminal voltage, Vt , electric power, Pg) indicates the magnitude and direction of the eigenvalue movement in the complex plane when a static controller is consideredable. Table5.4 contains the residues of transfer functions relevant in stabilizer design corresponding to the electromechanical mode. The phase of the residue informs about the phase compensation required at the eigenvalue frequency so the phase of the eigenvalue sensitivity becomes 1800 and the magnitude of the residue determines the gain required to achieve the desired damping. A speed stabilizer
j0.15
2200MVA
j0.5
E
Et P Q
j0.93
requires almost 900 of phase compensation whereas accelerating power or electric power stabilizers do not require phase compensation. The gain of the speed stabilizer will be greater than the gains of either accelerating or electric power stabilizers.
92
Table 5.1 Complex eigenvalues of SMIB with TCSC
Mode No. 1,2 3,4
Eigenvalue -13.494 17.304i -0.2576.772i
Frequency (Hz) 2.7541 1.0777
Damping %
61.5 3.8
Table 5.2 Real eigenvalues oF SMIB with TCSC Eigenvalue MODE TIME CONSTANT (S)
NO
5 6 7 8 9 10 11
14658.0 -1000.0 -78.9 -22.5 -1.9 -0.2 -1.0
0.0001 -0.0010 -0.0127 -0.0445 -0.5382 -5.0531 -1.0000
5.4.3 Analysis of two areas four machine with TCSC In this study, a two area interconnected four machine power system shown in Figure 5.6 is considered. The system consists of four machines arranged in two areas inter-connected by a weak tie line [14].
93
The location of TCSC is indicated in the diagram. The system is operating with area 1 exporting 400 MW to area 2. The equivalent circuit of the system is shown in Figure 5.6 Network and generator data can be found in [14]. We have assumed that the generators are equipped with static excitation systems and steam turbines. In this study, the TCSC is treated as a variable capacitance. Figure 5.7 contains a plot of the eigenvalues in the complex plane. Three pairs of poorly damped eigenvalues are found. They result to be associated to the rotor dynamics. The slowest eigenvalues are associated to the speed-governing systems whereas the fastest are associated to the excitation systems. The synchronous machine modes are in between. From the table 5.5, we see that the system is stable. There are four rotor angle modes. There mode shapes are described by the component of the right eigenvector corresponding to the generator speed
Egenvalue of two area four machine test system
8 6 4 2
Im ag
0 -2 -4 -6 -8 -80
-70
-60
-50
-40
Real
-30
-20
-10
Figure 5.7 Eigenvalue of two area test system
94
5.5 DESIGN OF TCSC POD CONTROLLER USING RESIDUE METHOD The uncontrolled system, Figure.5.7, has one inter-area oscillatory mode characterized by = -0.1211 j3.7559 with damping ratio = 3.22%. According to Table 5.7, the bus 8 has the largest residue and therefore the most effective location of the SVC and to apply the feedback control. Using the method presented in section 5.3, POD controller parameters are calculated in order to shift the real part of the oscillatory mode, to the left half complex plane. The obtained transfer function for the SVC POD controller is
1 10s 1+ 0.1329s 1+ 0.1329s HTCSC (s) = K * * * 1+ 0.1s 1+10s 1+ 0.4325s 1+ 0.4325s

K=
i d = 25.8963 Ri H1 (i )
Table 5.3 Eigenvector and normalized participation factor corresponding poor mode -0.29835+j7.8548
95
Right eigenvector -0.09 j0.55 0.012 j0.0015 0.016 j0.016 -0.028 j0.008 0.026+ j0.022 -0.047j0.07 -0.004 j0.00 0 0.7553 -0.029 j0.32 -0.0096 + 0.0003i
Left eigenvector -1.8*10-7 +j1.8*10-17 -214j2.25*10-16 -151 +j1.66 *10-15 5.81 + j0.0 34.58 +j45.61 34.58 -j45.61 0.16 -j0.067 0.16 +j0.067 0.12 + j0.0 -2.28*10-6 +j2.09*10-19 0
Participation factor 0.47159 0.47159 0.00959 0.01886 0.00292 0.02466 0.0006 0 0.00045 0.0001 0.00004
Participation state Machine angle
Machine speed 1
q-axis damper e q d-axis damper e d
q-axis damper e q
d-axis damper e d
Exiter v m Exiter v r1
Exiter v f x1 of Tcsc
x 2 of Tcsc
Table 5.4 Residue corresponding to local mode -0.257-j6.772
Transfer function
Residue -56.15133.27
Phase angle 67.150
Pg
-0.1252 + -25.044 0.0585i 0.0429 0.0087i -11.46
V1
96
Table 5.5 Complex Eigenvalue of Two Area Four Machines Test
Mode No. Complex Eigenvalue 1,2 3,4 5,6 7,8 9,10 11,12 13,14 15,16 17,18 -12.3267j 20.5784 -12.0224 j19.9823 -15.2167j15.8377 -14.8232 j 5.6141 -1.7779 j 6.4726 -1.9176j 6.7494 -0.11727j 3.6383 -5.1493j 0.04188 -0.07742j .22111
Frequency Damping ratio % (Hz) 0.08 99.99 0.08 99.99 0.53 97.97 0.43 1.20 1.16 0.60 0.10 0.09 -98.63 10.05 10.23 3.22 99.72 99.76
Table 5.6 Real Eigenvalue of two area four machine test
S/N Real eigenvalue Time constant 1 -1190.0316 0.0008 -1000.0716 2 0.0010 -999.9352 3 0.0010 -999.9232 4 0.0010 -999.9274 5 0.0010 -106.6204 6 0.0094 -108.4672 7 0.0092 -108.2704 8 0.0092 -107.7505 9 0.0093 -39.5747 10 0.0253 -39.0534 11 0.0256 -25.9974 12 0.0385 13 -24.4134 0.0410 14 0.0014 -714.2857 15 -0.0014 714.2857
97
Table 5.7 Participation of the generators in the electromechnical modes of the two area test system Mode Eigenval G1 G2 G3 G4 No. ue 9,10 -0.7647 0.0 0.0429 0.41007 0.5521 7.5680 11 6 i 45 -0.7514 0.4 0.0123 11 0.5364 2 11,12 7.3036i 04 2 0.02767 -0.1211 0.2 0.1419 0.34591 0.2434 46 9 4 13,14 3.7559i 13 Table 5.8 Sitting Indices of TCSC for two area four machine test
Mode residues of the transfer function P/kc

TCSC location
6-7 5-6 7-8 9-8 10-9 11-10
Normalised Residue |Ri| 0.001 0.735 1.000 0.181 0.000 0.788
5.6
SIMULATION RESULTS
The effectiveness of the proposed method of POD designed was tested on two- area four -machine systems. The analysis results for the two systems are presented in Tables 5.1 to 5.9
98
A three phase fault is applied for second test model at the bus 8 and cleared after 74ms. The original system is restored upon the fault clearance. The transient stability performances of the system with TCSC without POD and TCSC with POD controller are shown in Figures 5.8- 5.11. The TCSC with damping controller stabilizes system as can be seen from Figures 5.1-5.4. The oscillations of the system from Figures 5.1-5.4 are well damped with POD controller
Fault at bus 8
300 250 200 A tiv p w r in M c e o e W 150 100 50 0 -50 -100 -150 0 without POD controller with POD controller
10 Time(s)
15
20
25
Figure.5.8 Active power flow with and without POD in line 7-8
Fault at bu 8
20
10
Ag dvao ( 1G) n l e i t n G- 3 e
-10
-20
-30
-40
without POD with POD
-50
-60
10
12
14
16
18
20
Tim e(s)
Figure.5.9 Angle response of G1
Faut at bus8
50
99
R a tiv p w r inM a ec e o e Vr
-50
-100
-150
Without POD With POD
-200
Time (s)
10
12
14
16
18
20
Figure.5.10 Reactive power response for Bus 7

Fault is applied at bus 8
10
S e d d ia n in ra /s p e e tio d
-5
-10
-15
Without POD With POD 5 10 15 20
-20 0
Time(s)
Figure 5.11 speed response of G1
5.7 CONCLUSION This work has reviewed methods for analysis and control of power system oscillations with TCSC device based on the eigenstructure of the state matrix of the linear model of the power system. Residue-based methods also provide valuable information on how to design power system damping controllers. Although eigenvalue based methods are very powerful, the complexity of the power system stability problem requires the complementary use of other methods such as non-linear time domain simulation. All the simulations were done with PST toolbox in Matlab environment.
100
APPENDIX TCSC data Tr = 10 ms, XL = 0.2, XC=0.1,.Kc=50% Kr=10, TW = 10 s, MAX =3.1416, MIN =- 0.314 REFERENCES [1] B. Kalyan Kumar, S.N. Singh and S.C. Srivastava Placement of SVC controllers using modal controllability indices to damp out power system oscillations IET Gener. Transmission . Distrib., Vol. 1, No. 2, page 209-217, March 2007 [2] R J. Piwko, Ed., Applications of Static Var Systems for System Dynamic Performance IEEE Publication 87TH018755-PWR, [3] N.G. Hingorani, Flexible ac transmission, lEEE Spectrum, April 1993, pp. 40-45. [4] M. A. Pai and Alex Stankovic Robust control in power system tex book. USA, Springer Science+Business Media, Inc. 2005 [5] S.N. Singh, A.K. David A New Approach for Placement of FACTS Devices in Open Power Markets IEEE Power Engineering Review, 2001pp 58-60 [6] K.R. Padiyar Power system dynamics stability and Control Anshan limited UK 2004 [7] kamoto, H., Kurita, A., and Sekine, Y.: A method for identification of effective locations of variable impedance apparatus on enhancement of steady state stability in large scale power systems, IEEE Trans. Power Syst., 1995, 10, (3), pp. 14011407 [8] E. Acha, C. Fuerte-Esquivel, H. Ambriz-Perez and C. AngelesCamacho. FACTS Modelling and Simulation in Power Networks, John Wiley & Sons LTD, England, 2004. [9] C. A. Caizares and Z. T. Faur, Analysis of SVC and TCSC Controllers in Voltage Collapse , IEEE Trans. Power Systems, Vol. 14, No. 1, February 1999, pp. 158-165.
101
[10] G. Hingorani and L. Gyugyi, Understanding FACTS: Concepts and Technology of Flexible AC Transmission Systems, IEEE Press, 1999. [11] Federico Milano, Power System Analysis Toolbox Documentation for PSAT version 2.0.0 , March 8, 2007 [12] D.Jovcic, G.N.Pillai "Analytical Modelling of TCSC Dynamics" IEEE Transactions on Power Delivery, vol 20, Issue 2, April 2005, pp. 1097-1104 [13] Y.H. Song and A.T. Johns. Flexible AC Transmission Systems'. IEE Power and energy series, UK, 1999. [14] Kundur P. Power system stability and control. New York, USA: McGraw- Hill; 1994. [15] Bikash Pal , Balarko Chaudhuri Robust Control in power System Springer Science+Business Media, Inc.USA 2005 [16] J. Perez-Arriaga, G.C. Verghese, F.C. Schweppe, Selective Modal Analysis with Applications to Electric Power Systems. Part I:Heuristic Introduction. Pant 11: The Dynamic Stability Problem. IEEE Transactions on Power Apparatus , Vol. PAS-101, No. 9, September 1982, pp. 3117-3134 [17] Einar, V.L., and Joe, H.C.: Concepts for design of SVC controllers to damp power swings, IEEE Trans. Power Syst., 1995, 10, (2), pp. 948954
102
6
APPLICATION OF FACTS DEVICES FOR DAMPING OUT POWER SYSTEMS OSCILLATIONS USING MODEL TECHNIQUES
N. Magaji, M. W. Mustafa and A.S. Mokhtar
6.1 INTRODUCTION Variable impedance devices using power electronic technology, such as Static VAR Compensators (SVCs) Thyristor-Controlled Series Capacitors (TCSC) and Unified Power Flow Controller (UPFC) have the potential to increase power control and system damping [1, 2]. The allocation of these devices to get the most effective stabilisation of system modes is a complex problem that requires consideration of many factors. Major design considerations include the identification of busbars and branches in which to locate the system controllers and the choice of suitable feedback signals to be used as input to a stabilising control function [3-6]. Several analytical criteria for placement of damping controllers have been proposed in the literature. Specifically, modal analysis techniques based on the calculation of participation factors, transfer function residues and controllability and observability indices have been used to identify optimal locations for SVCs, TCSC, UPFC and power system stabilisers [2, 7-11]. The computation of controllability and observability measures is of particular interest as it provides valuable information about system dynamic characteristics and the ability of system controllers to enhance damping.
Application of FACTS Devices for Damping out Power
103
Systems Oscillations Using Model Techniques
This work presents an efficient analysis and design method to place variable impedance devices for enhancement of small signal stability in complex power systems. Controllability is first explored to allocate damping controllers in those busbars and system branches in which they are more effective to enhance damping of a given subset of critical oscillation modes. Observability indices are, in turn, taken to assess supplementary signals that are to be used in a given controller and which are required to show the oscillation mode. In this approach, the computation of controllability and observability is based on a measure of the distance of a state representation from the nearest uncontrollable system [12, 13]. This method is numerically stable and can be used to assess the numerical rank and other properties of state models. The study of inter-area oscillations and placement of controllers is done using modal analysis of a linearised model of the power system that includes a detailed representation of variable impedance controllers. The devices presently considered in this research include SVC, TCSC and UPFC. A reduced order equivalent of the interconnected New England test system (NETS) and New York power system (NYPS) is used to illustrate the proposed algorithms.
6.2
THE POWER SYSTEM MODEL
6.2.1 Formulation of the State Equation Multi-machine power system dynamic behavior in this frequency range is usually expressed as a set of non-linear differential and algebraic (DAE) equations. The high frequency network and stator transients are usually ignored when the analysis is focused on low frequency electromechanical oscillations. The initial operating state of the algebraic variables such as bus voltages and angles are obtained through a standard power flow solution [5].The state model proposed in [6] is adopted to represent system dynamic
104
behaviour. In this model, the power system is seen as constituted of dynamic subsystems interacting through the transmission network. Two main dynamic subsystems have so far been considered, namely synchronous machines and thyristor-controlled series capacitors. The dynamic behaviour of system is expressed in (DAE) as shown below after it has been Linearized around the equilibrium point.
x=f ( x,z,u )
0=g ( x,z,u )
(1) (2) (3)
y=h ( x,z,u )
Linearizing equations (1) to (3) around the equilibrium point {x0, z0 u0) g ives the following equations (4) to (6)
x =
0=
f f f x + z + u x z u
(4) (5) (6)
g g g x + z + u x z u
y=
h h h x + z + u x z u
Elimination of the vector algebraic variable z from equations (4) and (6), gaves
x = Ax + Bu y = C x + Du
(7)
where A is an system matrix, B and C are the column-vector input matrix and the row-vector output matrix, respectively. Let i = i + ji be the i-th eigenvalue of the state matrix A. The real part of the eigenvalues gives the damping, and the imaginary part gives the frequency of oscillation. The relative damping ratio is given by
105
2 + 2
(8)
If the state space matrix A has n distinct eigenvalues, and , and below are the diagonal matrix of eigenvalues and matrices of right and left eigenvectors, respectively:
A = A = = 1
(9)
In order to modify a mode of oscillation by feedback, the chosen input must excite the mode and it must also be visible in the chosen output. The measures of those two properties are the controllability and observability, respectively. The modal controllability and modal observability matrices are defined as follows:
Bm = 1 B Cm = C
(10)
The mode is uncontrollable if the corresponding row of the matrix Bm is zero. The mode is unobservable if the corresponding column of the matrix Cm is zero. If a mode is either uncontrollable or unobservable, feedback between the output and the input will have no effect on the mode. The open loop transfer function of a SISO (single input single output) system is:
G( s ) = y( s ) = C( sI A )1 B u( s )
(11)
G(s) can be expanded in partial fractions of the Laplace Transform of y in terms of C, B, matrices and the right and left eigenvectors as:
G( s ) = =
N
C i B i i =1 ( s i )
N
Ri ( s i ) i =1
(12)
Each term in the denominator, Ri, of the summation is a scalar called residue. The residue Ri of a particular mode i gives the
106
measure of that modes sensitivity to a feedback between the output y and the input u; it is the product of the modes observability and controllability. Figure 6.4 shows a system G(s) equipped with a feedback control H(s). When applying the feedback control, eigenvalues of the initial system G(s) are changed. It can be proven, that when the feedback control is applied, the shift of an eigenvalues can be calculated by (13) It can be observed from (19) that the shift of the eigenvalue caused by the controller is proportional to the magnitude of the corresponding residue. For a certain mode, the same type of feedback controls H(s), regardless of its structure and parameters can be tested at different locations. For the mode of the interest, residues at all locations have to be calculated. The largest residue then indicates the most effective location to apply the feedback control.
i = Ri H( i )
6.3 FACTS MODELS
Figure 6.1. Modelling of FACTS Devices.
107 is
In addition, the interconnecting represented by the node equation:

i = YN V + TN x
transmission
network (14)
With:
x = [x d1 ,.., x dm , x tsc ] T , V = v D , v R , VD = [ v t1 ,.., v t m ] , VR = [ v r1 , .., v
i = [i d1 ,.., i dm ,0] T
rk
Where x is the system state vector and xtsc is the state vector associated with the dynamic representation of TCSC. Matrix YN is the network admittance matrix and TN appears from the representation of series-connected variable impedance devices, namely TCSC and thyristor controlled phase shifters. Static loads are taken into account as non-linear functions of their terminal voltage changes and introduced into the admittance matrix . 6.3.1 The Overall State Model The overall state model is obtained using a component connection model. The composite model for the system is given by:
m x = A x + b ju j j =1 yk = Ck x
(15)
where uj is the control signal to the j-th variable impedance device and yk is the k-th measurement signal. The state matrix Anxn is defined symbolically as:
A = A D + CD [ YN YD ]
1
[TN + WD ]
(16)
System eigenvalues are calculated using an algorithm based on the Arnoldi's method for computing eigen-elements of large unsymmetrical matrices [15] Let the power system dynamic behaviour be given by the linear state representation in (4) where xRn denotes the state vector,
108
uRm the input or control vector for B=ctrbl[b1,.., bm], and (A,B) Rnxn x Rnxm with mn. The pair (A,B) is controllable if and only if rank [I-A,B]=n for each eigenvalue of A [12]. Similarly, the system in (7) is completely observable if and only if rank [I-AT, CT]=n for each eigenvalue of A. 6.4 A MEASURE OF MODAL CONTROLLABILITY AND OBSERVABILITY A practical measure of controllability is obtained from the distance of the pair (A,B) to the nearest uncontrollable system [12,13]. Let the set of systems SYS having n states and m inputs be defined as:
SYS n ,m =
{( A, B)
A C nxn , B C nxm
(17)
Then let the set of uncontrollable system UNCOn,m in (A,B) be defined as:
UNCOn ,m =
{ ( F , G ) ( F , G) SYS n,m , C: rank [I F , G ] < n }
(18)
The distance between a system (A,B) and the set of uncontrollable systems (F,G) is given by the following theorem [13] . Theorem. The distance d between the pair (A,B)SYSn,m and the set of uncontrollable systems is
d ( A, B ), UNCOn ,m = min min [I A, B] for C
(19)
where min is the minimum singular value of matrix [I-A,B]. A dual definition can be stated for the analysis of observability. The minimum singular value of the augmented matrix Pij=[iIA,bj] provides a measure of the distance from the nearest singular system. This can be interpreted as the degree of coupling of the j-th input to the controller (the k-th measurement signal from the system) to the i-th eigenvalue [12].
109
6.4.1 Computational algorithm The allocation algorithm can be formulated as follows: a) Determine the set of critical eigenvalues i (i=1,..,p) b) Calculate the coupling indices Pij and Qik for each of the selected compensation alternatives (busbars and branches) and input signals. c) Calculate the singular value decomposition of matrices Pij and Qik. The minimum singular values provide a quantitative measure of the effect of system controllers on system damping. d) Define a location index (RI.) as the product of controllability and observability measures e) Rank busbars and branches according to the product of controllability and observability indices 6.5 CALCULATION FOR MODAL CONTROLLABILITY, OBSERVABILITY AND RESIDUE 6.5.1 Dynamic Model of the System Each generator of the test systems is equipped with the AVR Type III which is the simplest AVR model that can be used for rough stability evaluations [13].The generator is described by six order non-linear mathematical model while exciter by third order [13]: The sixth order model of generator is obtained assuming the presence of a field circuit and an additional circuit along the d-axis and two additional circuits along the q-axis. The generator state variables are ( , , eq , ed , eq and ed ) while exciter has the following state variables (vm , vr , v f ) .The state variables are also defined in M.A Pai [21]. (20) (21) (22) Y = C X + D U Where A, B, P, Q, C, and D are the matrix of partial derivatives in (13) to (15) evaluated at equilibrium points. In PSAT those
X = A X + B U 0 = P X + Q U
110
equilibrium points or initials conditions are obtained after power flow simulation of the test system. Where X = [i , i , e , e , e , e , v mi , v ri , vfi ] (23) di di qi qi (24) U = [Idi , Iqi , Idid , i , Vi , k , Vk ] Iq and Id are the quadrature and direct axis components of the stator current, respectively, while V is the complex bus voltage. The order of X is k=9m+n where m is number of generators and n is the order of the FACTS device with the six order machine model and third order exciter dynamic components. m=10 and 12 in test system1 and system 2 respectively while n =1 for SVC 2 for TCSC (Reactance Model) and 3 for UPFC in both cases. 6.5.2 Residue, Controllability and Observability for SVC Placement Let an SVC, with susceptance Bsvci , be placed at bus i. The reactive power injected into the bus (Qsvci) is a function of SVC susceptance Qsvci = BSVCV i2 (25) Where Qsvci and Vi are the reactive power injected and voltage at bus i, respectively. Considering initial value of Bsvci as zero and linearizing (25), we obtain Qsvci = V i2 BSVC (26) Combining (21) and (26) provides P X + Q U + LBsvc = 0 (27) Where matrix L contains partial derivatives of power balance equations at all the buses with respect to the susceptance, provided by the SVC, initially assumed to be placed at all the buses. Solving (20) and (27), we obtain X= ( A-BQ -1 P ) X+(-BQ -1 L)Bsvc (28) Equation (28) can be rewritten as X=AX+LBsvc (29) Where w is the left eigen vector of matrix A corresponding to the kth mode.
111
The controllability index of SVC placed at the ith bus to the kth mode can be defined as given in equation (7) CIi = w T L (30) k i To get the observability index of SVC, substitute equations (26) in (22) Y= ( C-DQ -1 P ) X+ ( -DQ -1 L ) Bsvc (31) Therefore equation (31) can be written as Y=C X+DBsvc (32) Where E is an output matrix with SVC device. The observability index of SVC placed at the ith bus to the kth mode can be defined as given in equation (10) OIi = Cv k (33) i Therefore the product of equation (30) and (33) will give the Residue factor defined by equation (12)
6.5.3 Residue, Controllability and Observability for TCSC placement Consider a line l, having line reactance XL, connected between buses k and m. If the reactance of TCSC placed in the line l is Xc, the percentage of compensation of TCSC (kc) is given by X kc = C (34) XL The line power flows are functions of the degree of compensation of the TCSC. The real power (Pkm) and reactive power (Qkm) in a line l (connected between buses k and m), with TCSC having degree of compensation Kc and neglecting the line resistance, can be written as [11]
Pkm = Kc B km V k V m sin ( k - m ) Kc 1
(35)
Qkm =
Kc Bkm Vk2 -VkVm cos( k - m ) Kc 1
(36)
112
Where Bkm is the susceptance of the line l, Vk k is the complex voltage at bus K. Considering the initial value of Kc to be zero, (35) and (36) can be Linearized, around Kc, Pkm = K cV k V m B km sin ( k - m ) (37) 2 Qkm = K c Bkm (Vk -VkVm cos( k - m )) (38) Combining (35) and (36) with (21) , we obtain P X + Q U + F K c = 0 (39) Where matrix F contains partial derivatives of power balance equation at all the buses with respect to degree of compensation, provided by TCSC, initially assumed to be placed in all the lines. The order of matrix F is (2m +2n) X nl, where nl is the number of lines Solving equations (21), (22) and (39) provides (40) X= ( A-BQ -1 P ) X+(-BQ -1 F)K c
Y= E-DQ -1 P X+(-DQ -1 F)K c
(41)
Equations (40) and (41) can be rewritten as (42) Y=C X+D K c (43) The controllability index of the TCSC, placed in line l, to the kth mode is defined CIi = w T Fi (44) k th While the observability index of the TCSC in line l, to the k mode is defined (45) O Ii = C v k i Where w and v are the left and right eigen vector of matrix A corresponding to the kith mode respectively. Therefore the product of equations (44) and (45) will give the Residue factor defined by equation (12)
X=AX+F K c
6.5.4 Residue, controllability and observability for UPFC TCSC placement Consider a UPFC placed in a line l, connected between buses i and j. UPFC can be represented by one series voltage source VS and by
113
one shunt current source ISH, The series voltage source and the shunt current source are defined as follows: VS =Vp + jVq = Vs e j (46) j I SH =I p + jIq = I e (47) Assuming bus i as the sending end (UPFC placed towards the bus i and bus j as receiving end, connected in a line l. The series injected voltage can be separated into two components, in phase (Vp) and in quadrature (Vq) to the sending bus voltage. Vq has relatively strong effect on real power and Vp on the reactive power flow in the line l. Thus, Vq can improve the damping of inter-area oscillations [4]. Taking Vq as the control variable for UPFC, real power flow in the line l can be written Pij =ViV j bij sin ( i - j ) +VqVi bij (48) Real power in line l, without UPFC, is Pij =ViV j bij sin ( i - j ) (49) Subtracting (49) from (48), we obtain the expression for additional real power injected owing to UPFC (Pij_upfc) (50) Pij_ upfc =VqVi bij Linearizing (50), with initial value of Vq considered as zero, we obtain Pij_ upfc =VqVi bij (51) Combining equations (51) with (21) provides PX +QU +G Vq =0 (52) where matrix G contains partial derivatives of the power balance equation at all the buses with respect to the series voltage injection of UPFC, initially assumed to be placed in all the lines. The order of matrix G is (2m + 2n)X(nl). Solving equations (20) and (52), we obtain X=(A- BQ -1 P)X +(-BQ -1G)Vq (53) Equation (53) can be rewritten X = AX + G Vq (54) The controllability index of UPFC for series branch placed in line l to the kth mode is defined as
K
114
CIseries (55) To get the Observability index of UPFC Vq series voltage, substitute equations (52) in (22) Y= ( C-DQ -1 P ) X+ ( -DQ -1G ) Vq (56) Therefore equation (56) can be written as (57) Y=C X+DVq Like wise the Observability index of UPFC for series branch placed in line l to the kth mode is defined as OIseries OI iseries = Civk (58) Therefore the product of equations (55) and (58) will give the Residue factor for series UPFC defined by equation (12). Since UPFC is a series-shunt device, the selection of the bus (end of the line), where UPFC shunt converter should be connected, is also important. If the shunt branch of UPFC injects reactive current (Iqi) at bus i, which is in quadrature to the voltage at bus i, the reactive power injected into the bus can be written Qi =I qiVi (59) Considering initial value of Iqi as zero and linearizing (52), we obtain Qi =Vi I qi (60) Combining (53) with (14) provides P X +QU+H Iq = 0 (61) where matrix H contains partial derivatives of power balance equation at all the buses with respect to reactive current injection of shunt branch of UPFC. The order of matrix H is (2m+2n) X (2). Solving equations (54) and (13), we obtain X=(A- BQ -1 P)X +(-BQ -1 H)I q (62) Equation (55) can be rewritten as X = AX + H I q (63) The controllability index for shunt branch of UPFC placed at the ith bus to the kth mode can be defined T CI ishunt = vk H i (64)
T CI lseries = vk Gl
115
where v is the left eigen vector of matrix A corresponding to the kth mode. To get the Observability index of UPFC Iqi currents, substitute equations (54) in (15) Y= ( C-DQ -1 P ) X+ ( -DQ -1 H ) I q (65) Therefore equation (49) can be written as Y=C X+DI q (66) Like wise the Observability index of UPFC for series branch placed in line l to the kth mode is defined as OIishunt OI ishunt = Civk (67) Therefore the product of equations (57) and (51) will give the Residue factor for shunt UPFC defined by equation (9).
6.6 APPLICATION EXAMPLE The proposed methodology is tested on a dynamic model of the reduced order equivalent of the interconnected New England test system (NETS) and New York power system (NYPS) Generators G1G9 are the equivalent representation of the NETS generators, and G10G13 represent the equivalent generators in the NYPS. Generators G14 to G16 are the dynamic equivalents of the three neighboring areas connected to the NYPS. The test system consists of 68 buses, 16 machines, and five areas .The static and dynamic data are adopted from [16].
Modal analysis shows that, in the base case, there is an interarea mode with low damping ratio of 0.007085, having eigenvalues 0.02322 j3.2774. Analysis of phase angle of the right eigen vectors shows that in this inter-area mode, generator G15, which is in a neighborhood of NYPS, is oscillating against a cluster of generators G13 in NYPS, G9 and G6 in NETS. For this critical mode, the controllability indices of all the two types of FACTS controllers for the base case were determined and given in the normalized form in Table 6.1. From Table 6.1 it can be observed
116
that the lines between buses 5152 and 26-27 are the optimal locations for TCSC and UPFC respectively, while bus 28 is the best location of SVC device
117

Table 6.1 Residue Factor of FACTS Device for 68-bus test system TCSC Placement SVC Placement UPFC Placement
Line connection 2 3 2 25 10 11 10 13 16 19 21 22 23 59 9 36 36 37 34 36 44 45 45 50 51 52 50 51 41 40
Bus No. N RI 0.17 0.27 0.1 0.12 0.21 0.23 0.47 0.25 0.63 0.75 0.11 0.45 1 0.19 4 5 6 8 11 14 16 18 28 39 40 43 45 49
N RI 0.089 0.09 0.089 0.096 0.082 0.084 0.078 0.083 1 0.178 0.398 0.158 0.29 0.244
Line connection 2 3 3 18 9 30 16 24 21 22 26 27 36 37 34 36 33 34 1 31 31 38 1 47 48 40 35 45 44 45
N RI 0.15 0.21 0.05 0.98 0.35 1 0.03 0.03 0.11 0.17 0.55 0.49 0.06 0.03 0.08
0.15 51 0.131 Where NRI= Normalized Residue Index
6.7 CONCLUSIONS A set of Residue indices have been proposed in this work in order to find the optimal location of UPFC, TCSC and SVC for enhancing damping of small signal oscillations and, thus, improving the system stability. Residue indices for the placement of FACTS controllers have been found corresponding to a critical inter-area mode in case of 68-bus systems. These indices have been computed for the base case only Application studies on a complex power system have shown that SVCs TCSC and UPFC have the potential to substantially increase
118
system damping. Further refinements are however required to discriminate the impact of several compensation alternatives.
REFERENCES [1] A. R Messina and N. M., "An Efficient Placement Algorithm Of Multiple Controllers For Damping Power System Oscillations," Power Engineering Society Summer Meeting, 1999. IEEE, vol. 2, pp. 1280-1285., 1999. [2] IEEE FACTS Working Group, FACTS Applications, IEEE Publication 96TP116-0, 1995. [3] Current Activity in Flexible AC Transmission Systems, IEEE Publication 92TH 0465-5 PWR, April 1992. [4] B. Kalyan Kumar, S.N. Singh, and S. C. Srivastava., "Placement of FACTS controllers using modal controllability indices to damp out power system oscillations," IET Generation. Transmission and Distribution., vol. 1, pp. 209 217., 2007. [5] Larsen, E.V. and Chow, J.H. SVC Control design Concepts for System Dynamic Performance. Application of Static VAR Systems for System Dynamic Performance, IEEE Tutorial Course 87TH0187-5-PWR, 1987. [6] Grund, Carl. E, Larne Einar, V., Murdoch, Alexander, and Paserba, John, J. Mitigation of Inter-Area Oscillations by Control. IEEE Winter Meeting, Symposium on Inter-area Oscillations, New York, NY, February, 1991. [7] Larsen, Einar, V. and Chow, Joe, H., Concepts for Design of FACTS Controllers to Damp Power Swings. IEEE/PES 1994 Summer Meeting, San Francisco, CA. July, 1994, paper 94 SM 532-2 PWRS. [8] Clark, K., Fardanesh, B., and Adapa, R. Thyristor Controlled Series Compensation Application Study-Control Interaction Considerations.IEEE/PES 1994 Summer Meeting, San Fco., July 1994, paper 94 SM 478-8 PWRD.
119
Yang, Xiaoqing, Feliachi, Ali. Stabilization of Inter-area Oscillation Modes Through Excitation Systems. IEEE Transactions on Power Systems, Vol. 9, No.1, Feb 1994, pages 494-502. [10] Martins, Nelson, and Lima, Leonardo. Determination of Suitable Locations For Power System Stabilizers and Static VAR Compensators for Damping Electromechanical Oscillations in Large Scale Power Systems. IEEE Transactions on Power Systems. Vol. 5, No.4, Nov 1990, pages 1455-1469. [11] Okamoto, Hiroshi, Kurita, Atsushi, Sekine, Yasuji. A Method for Identification of effective Locations of Variable Impedance Apparatus on Enhancement of Steady-State Stability in Large Scale Power Systems. IEEE Transactions on Power Systems. Vol. 10, No.3, August 1995, pages 1401-1407. [12] Vrouco L.,and Pagola, F.L. An Eigenvalue Sensitivity Approach to Location and Controller Design of Controllable Series Capacitors for Damping Power System Oscillations. IEEE/PES Winter Meeting, New York, Feb 1997, pages 14551458, paper PE-547-PWRS-0-01-1997. [13] Nayebzadeh, M., and Haubrich, H.J. Placement of Static Var compensators for small-signal stability enhancement. Colloquium of CIGRE Study Committee 38, Florianopolis, Brazil, Sep 1993, pages 1-4. [14] Paige, Chris C. Properties of numerical algorithms related to computing controllability. IEEE Transactions on Automatic Control. Vol. AC-26, No.1, Feb 1981, pages 130-138. [15] Eising, Rikus Between Controllable and Uncontrollable. Systems and Control Letters. North Holland, 4, July 1984, pages 263-264. [16] Romn Messina, A., and Cory, B. Enhancement of Dynamic Stability by Co-ordinated Control of Static VAR Compensators. Int. Journal of Electrical Power & Energy Systems. Butterworth-Heinemann Ltd, Vol. 15, 4(1993), pages 85-93.
[9]
120
Scott, J.A. An Arnoldi Code for Computing Selected Eigenvalues of Sparse Real Unsymmetric Matrices. Rutherford Appleton Laboratory Report RAL-93-097 (1993). [18] Romn Messina, A, Ramrez, J.M., and Caedo, J.M. An Investigation on the Use of Power System Stabilizer for Damping Inter-Area Oscillations in Longitudinal Power Systems. IEEE/PES Winter Meeting, New York, NY, Feb 1997, Paper 94SM 532-2 PWRS.
[17]
Index
121
INDEX
Adaptive contro,l, 2 Accelerating power, 4 AVR, 133 controllability, 24, 21,26,27, 43,45,47,48,64,65,100,103 ,104,125,126,128,131,132, 135,136,139 Controllers, 11, damping controller, 2,39,42,57 120 damping ratio, 8, 38 53 damping oscillations, 1,7,24,30,68,70,89 differential and algebraic (DAE) , 32,52,91,116 eigenvalue sensitivity, 103,133 eigenvalues, 22,24, 25,29,31,132 excitation system, 1,9,40,60,75,101,103, 105,
FACTS devices, 1,30,38,68,69,70,90,115 feedback signal, 31,51,65,70,76,90,115, 120 Hankel singular values , 65,76 HSV, 65,76 intelligent control, 8 interarea oscillations, 11 local mode,7,11,14,44 LQR, 9,10,17,19,21,25,26. Modal controllability, 34, 66, 93,112,118,121,122 mode shape, 33,61,92,105 Multi-machine, 32, 52,91, 116 Negative damping, 26
122
Index
observability, 118 observability index, 1,3,35,95,126,127,128,129 positive damping, 24 power system stabilizers, 1,23,31,51,68,90,102. power systems, 1,52,89,90,97,101,115,116. residue, 76,81,84,90,96,103,106,111,1 15, 119,122,126,129,131,134, Residue factor, 1, 50,71,76,84,124,126,128, 129, right eigenvectors, 32,52,76,81,91 robustness, 1,23 singular values, 122 SISO, 36,55,71,95,118
small signal stability, 7,116. SMIB, 7,8,10,12 transient stability, 96,97,110 torsional mode, 7 uncertainty, 24 washout,11,12, 38,59,100

Bookchapter Fke40

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Bookchapter Fke40

Caricato da

Copyright:

Formati disponibili

First Edition 2008

AZHAR KHAIRUDIN 2008

Perpustakaan Negara Malaysia

Diterbitkan di Malaysia oleh / Published in Malaysia by

Azhar Khairuddin Faculty of Electrical Engineering Universiti Teknologi Malaysia 2008

Developments of Power System Dynamic for Damping Oscillation

K3 K3 K 4 E fd ' 1 + sTd 0 K 3 1 + sTd' 0 K 3

Developments of Power System Dynamic for Damping Oscillation

Developments of Power System Dynamic for Damping Oscillation

Figure 1.2 : Linearized model of SMIB system

Figure.1.3 : Linearized model of SMIB system

Developments of Power System Dynamic for Damping Oscillation

PROPOSED LQR POWER SYSTEM STABILIZER ALGORITHM

Developments of Power System Dynamic for Damping Oscillation

R =T T Where T is a nonsingular matrix And

[TK (T T )1 BT P ] PBR 1BT + Q = 0

(24) (25) (26)

Thus we get a control law as

Weight Matrix Selection

Developments of Power System Dynamic for Damping Oscillation

The dominant poles -13.6367 j30.0225 can be rewritten as:

Light load DW response

Conventional PSS NO PSS LQR PSS

Time in sampling of 0.1

Figure.1.4 Vs time for light load condition of the SMIB system

0.01 0.005 0 -0.005 -0.01 -0.015 -0.02 -0.025 0 20 40 60 80

N0 PSS Conventional PSS LQR PSS Time in 0.1 sampling

-0.01 -0.02 -0.03 -0.04 -0.05 0 10 20 30 40 50 60 70 80 90 100

Time in 0.1 sampling

Figure.1.6 Vs time for Heavy load condition of the SMIB system

Developments of Power System Dynamic for Damping Oscillation

Light load DW response

Conventional PSS NO PSS LQR PSS

Time in sampling of 0.1

Time 0.1 sampling

Figure1..8 Vs time for Normal load condition of the SMIB system

Time in 0.1 Samping time

Table 1.1 Operating conditions

Case 1 Normal 2 Heavy 3 Light

P(pu) 1.0 1.2 0.2

Q(pu) 0.1 0.2 0.05

Table 1.2 Eigenvalues

-12.7 j34.8 -10.5j20.8

Table 1.3 Controller parameters CONVENTIONAL PSS LQR Constants

7.7 0.05 0.05 0.2287 0.2287

Normal 0.6292 0.8050 0.8413 0.0033

Heavy 0.5407 9.4519 0.8061 0.0032

Light 0.6315 5.5692 0.2576 0.0013

Developments of Power System Dynamic for Damping Oscillation

1.5 COMPARISON OF VARIOUS DESIGN TECHNIQUES

Developments of Power System Dynamic for Damping Oscillation

(29) (30) (31) (32) (33) Eq 0 E0 cos 0 xe + xq (34)

Eq 0 = vq 0 + id 0 xq E0 = (vd 0 + xeiq 0 ) 2 + (vq 0 xe id 0 ) 2

(35) (36) (37)

xq vd0 xd' vq0 K5 = E0 cos0 E0 sin0 ' xe + xq V0 xe + xd V0 t t

Developments of Power System Dynamic for Damping Oscillation

Developments of Power System Dynamic for Damping Oscillation

Pre-multiplying both sides of (3) by V-1 gives

= diag ( 1 2 ... n-1 , n )

(4) (5) (6) (7)

A similar expression holds for the left eigenvectors W such that

Developments of Power System Dynamic for Damping Oscillation

Modal controllability and observability factors