First International Power and Energy Coference PECon 2006 November 28-29, 2006, Putrajaya, Malaysia

202

A Novel Approach to Optimal Allocation of SVC using Genetic Algorithms and Continuation Power Flow
S.R.Najafi

M.Abedi

S.H.hosseinian

Department of Electrical Engineering

Amirkabir University of Technology

Tehran,Iran

Abstract - This paper proposed a new approach to determine

optimal location of SVC to improve voltage profile and maximizing system loadability with and without generators Mvar limits. A variable reactance model for SVC at steady state studies are presented and implemented in load flow program with embedded FACTS Devices. A simultaneous GA and CPF used to determine maximum number of SVC and steady state stability margin, based on closing to point of voltage collapse. As an important result in this paper we obtained a maximum number of SVC beyond which system loadability can not be increase and hence increasing loading level leads to static voltage collapse phenomena. A case study and simulation are done on IEEE57 Bus Test System. In Genetic Algorithm optimization procedure the system loadability and bus voltage profile flattening are used as measuring of power system performance Index.

Index terms - Bifurcation Theory, CPF, Genetic Algorithms, Loadability, SVC, Stability Margin, Voltage Collapse.

II. SVC Steady state model There are several methods to describe SVC model in steady state and power flow studies, the power injection model and representing SVC as a controlled variable reactance. In these condition controller represent as a reactive power source and usually control limits applied on reactive power.[5], [6], [7], [8]. The current paper concentrates on describing an adequate SVC models, based on firing angle criterion. It is demonstrated here that the proposed model allow representing SVC in accurately and reliably way, for power flow and voltage stability studies. A schematic diagram of SVC illustrated in Fig. 1 and The basic structure of an SVC operating under typical bus voltage control is depicted in the block diagram of Fig. 2. Assuming balanced, fundamental frequency operation, an adequate transient stability model can be developed assuming sinusoidal voltages.

I. INTRODUCTION Increased loading of power system, environmental restrictions, combined with a world wide deregulation of the power industry, require more effective and efficient control means for power flow and stability control. The power flow control and static stability limits of power system can be considerably modified by the new reactive compensation equipment [1], [2].The development and use of Flexible AC Transmission System (FACTS) controllers in power transmission systems has led to optimal placement and applications of these controllers to improve bus voltage profile, stability of power networks and increasing power system loadability [3]. Many studies have been carried out and reported in the literature on the use of these controllers to voltage and angle stability applications. A variety of optimizing techniques are used to optimal allocation of FACTS devices in power system. In [4] the author's used GA to optimal placement of multiple choice FACTS controller to improve bus voltage profile. In mentioned paper FACTS devices are describe with simplified reactance model and the fitness function only involved bus voltage flattening and lines loading level. In this paper an exact modeling of SVC at steady state studies are presented and implemented in load flow program. The SVC is a shunt connected reactive compensation equipment which is capable of generating reactive power whose output can be varied so as to maintain control of specific parameters of the electric power system[3], [4], [5].

Tl

i
c

-L

Fig. 1: schematic diagram of SVC

This model can be represented by the following set of p.u. deferential and algebraic equations: [6], [8]:
(c X VREF

VB)

(1)
(2)
(3) (4)

Vk- - VREF + XSLOPE V BEq QSVC - V2 BE =0

ZVXCXLBE + sin 2cx- 2a +TK2 -XL) =0

1-4244-0273-5/06/$20.00 C2006 IEEE

203
All of the variables

are

indicated in

Fig.

and

Fig.

2.

XV

MAGNUDE

Be-

t)

Fig. 2: Transient stability model of a SVC

IV. Loading Parameter and CPF The most accepted analytical tool used to investigate voltage collapse phenomena is the bifurcation theory, which is a general mathematical theory able to classify instabilities, studies the system behavior in the neighborhood of collapse or unstable points and gives quantitative information on remedial actions to avoid critical conditions [11]. In the bifurcation theory, it is assumed that system equations depend on a set of parameters together with state variables, as follows: X = f(X,2) (5) Where X represent power system state variable and A is loading parameter. Then stability or instability properties are assessed varying "slowly" the parameters. The parameter that used to investigate system proximity to voltage collapse is called loading parameter A, which modifies load powers as equations (6) and (7): P = (1 + k12)Po (6) Q = (1+ k22)QO (7) Where P0 and Q0 are the active and reactive power at basic operating point at buses and k1,k2 are the control parameters to increase bus loading level. In typical bifurcations diagrams voltages are plotted as functions of i.e. the measure of the system loadability, called P-V or nose curves. Equations (6) and (7) used in continuation power flow analysis.
2,

Ff0
I1
O.

J]1

|Fz

1,]m m

Ii

Fig. 3: Crossover Operator

I
!--I

B. Mutation In natural evolution, mutation is a random process where one allele of a gene is replaced by another to produce a new genetic structure. In GAs, mutation is randomly applied with low probability, typically in the range 0.001 and 0.01, and modifies elements in the chromosomes. Fig. 4 shows the mutation process in genetic algorithms.
1-.'

10111 4j01

;_llo

_ 1O1 1(1i1 ml

_oli

A. Continuation Power Flow (CPF) Continuation Power Flow techniques are widely recognized as a valuable tool to determine nose curves of power systems and allow estimating the maximum loading conditions and "critical" solutions (for instance, saddle-node and limit induced bifurcation points). Although computationally demanding for large systems [12],[13], CPF is not affected by numerical instabilities ( it is able to determine the stable and unstable fold of P-V curves) and can provide additional information, such as sensitivity factors of the current solution with respect to relevant parameters [14]. From a mathematical point of view, the CPF is a homotopy technique [2] and allows exploring stability of power system equations when varying a system parameter, which, in typical static and dynamic voltage stability studies, is the loading parameter [15], [16], [17] and [18]. Generally speaking, CPF consists in a predictor step realized by the computation of the tangent vector and a corrector step that can be obtained either by means of a local parametrization or a perpendicular intersection [19].
V. SIMILATION PROCESS AND RESULTS

Fig. 4: Mutation Operator

C. Selection Selection is the process of determining the number of times, or trials, a particular individual are chosen for reproduction and, thus, the number of offspring that an individual will produce. The selection of individuals to produce successive generations plays an extremely important role in a genetic algorithm. A probabilistic selection is performed based upon the individuals. Individuals that have the better Fitness have an increased chance of being selected [9], [10].

In this section for the clarity of our proposed technique the flowchart for combination of GA, CPF, predictor and corrector is illustrated in Fig. 5. Also both Genetic Algorithms and Continuation Power Flow are time consuming algorithms in simulation process, since our studies are in the planning stage; hence the accuracy and efficiency of results are more important than time speed. In the optimization problem, we are concerned with the attempt to minimize the voltage deviation equation (8) and to maximize the system stability margin (SM) and increasing maximum system loadability equation (10) simultaneously

204

in IEEE57 Bus Test System (Fig. 10). Maximizing the system security margin Involves designing a system that is able to tolerate more load demand without voltage collapse.
Bus number

FV =
Where;

j7
P=1

Vp

(8)

exp{[abs(I- Vi) dV]} (9) Which Vi represents the voltage of bus i and dV is the deviation from best case. An objective function to maximize the stability margin and increase maximum system loadability obtained from equation (10)
(10) In equation (10) SM is power system stability margin. When loading parameter increase, stability margin will close to number one.
1

VP

F(X, Y, A)

The fitness function consists of two optimization problem. Equation (9) defines voltage deviation and Equation (10) describe power system stability margin to voltage collapse. With optimal placement of SVC, it is possible to increase the system loadability. We found a maximum number of device beyond which the efficiency of the network cannot be further improved. According to the optimization criterion, for the tested power system, the results show that this limit is about 10 devices for system with generator Mvar limit and about 7 for system without generator Mvar limits. Fig. 6 shows loading parameter i, against SVC number with and without generators Mvar limit. It should be notice that Fig.6 is a discrete response. A glance to Fig. 6 indicates that maximum suitable number in case with generators Mvar limit is about number 7 and without generator Mvar limits the maximum suitable number in this case is about number 10.
i~

~ ~

MAXIMUM LOADING PARAMETER (AMEDA)

17

l
0

...T

,,. .,j,,,,., TS.,,i

uJ LU

< 15
14

I*
* 9

GEEAO LIMIS

P,

1AS

<

1.2
1.1

YES

16

II

10

15

20

25

so

35

40

45

SEV NUMBER

No

Fig. 6: Number of SVC vs loading parameter (with and without generators Mvar limit)

CROSSOVE1

YES

I NO

Fig. 5: Fitness Function Evaluation Flowchart

From Fig. 6 After optimal allocation of SVC maximum = 1 to system loadability increased from base case = 1.63 with generators MVAR limits and it increased to is A = 1.78 without generator MVAR limits, which loading parameter. Results show that maximum system loadability with generators MVAR limits is smaller respect to case without generator MVAR limits. Fig. 7 show GA trace for placement of SVC to maximize system loadability and improving bus voltage profiles. The GA simulation after 100 generation was reached to optimal solution. In Fig. 8 we have shown P-V curves that indicate power system trajectory leading to stability margin and closing to voltage collapse phenomena in presence of SVC and without it. Fig. 8 shows that with SVC installation maximum loadability of power system considerably increase and static stability margin improved. In Fig. 9 all bus voltages profile are shown before and after SVC placement in mentioned network. Results indicate that after SVC allocation we have a flatter voltage profile moreover optimal placement of shunt FACTS Devices in IEEE57 Bus unstable or stressed system not only increase system loadability but also prevent to voltage collapse.

To simulation and test of model a program are developed in MATLAB environment and the following results obtained.

205
Performance Index

.2
0 (A 0 r1--

r-

t r_

The location and number of SVC's obtained by GA and the value of SVC'Mvar determined by continuation power flow. Considering tablel and Fig. 10 we find that almost all SVC's are installed at weak bus from reactive power standpoint. A glance to Fig. 10 indicates that these buses are for from reactive power source and hence are critical in voltage phenomena.
TABLE 1: SVC placement (bus-Mvar-number) 15 31 36 42 52 Bus number Mvar 50
56

7.8

11.28

15.3

5.49

3.76

10

20

30

40 50 60 Generation Counter

70

00

90

100

Fig. 7: GA trace for placement of SVC to maximize system loadability

*S

16

8U8

II
_6+1

1-

"I

B36

a C 1
0)

29

jj28

-\ 27
51

27

X
H2

Fg1:IE5BsTsSs-

Fig. 10: IEEE 57 Bus Test System

Fig. 8: P-V comparison curve for two case (with and without SVC allocation)
1.4

BUS VOLTAGE

1.2 F

0.x

0.6

mm0.
i!

{LIe

~ -

~~~~~

without ~ ~~~~ Rh

0.4 F

modeled for steady-state and transient stability studies. Optimizations were performed on three parameters, the number of the device, their location, and their value. The system loadability and improving bus voltage profile was employed as measure of power system performance in optimization algorithm. Simulations are down in both cases with generator Mvar limits and without generator Mvar limits. For any cases power system loadability was increased but this value with generator Mvar limits was smaller, Moreover Results have shown that the simultaneous use of GA and CPF are efficient solution to find the maximum system loadability and optimal placement of SVC. We observed a maximum number of SVC beyond which the loadability cannot be improved. Besides this power system elements are located at buses that are weak from reactive power requirements.
VII. REFERENCES
[1] E.Acha, V.G.Agelidis, O.Anaya and T.J.Miller, Power Electronic Control in Electrical System, Newness power engineering series. First published 2002. [2] Modeling of Power Electronics Equipment (FACTS) in Load Flow and Stability Programs: A Representation Guide for Power System Planning and Analysis, Technical brochure no. 145, TF 38-01-08, CIGRE (Aug. 1999). [3] N.G.Hingorani and .Gyugyi, Understanding FACTS: Concepts and Technology Of Flexible AC Transmission System, Institute of Electrical and Electronics Engineers, New York, 2000.

VI. CONCLUSION We have presented simultaneously GA and CPF to optimal allocation of SVC in a power system. This device is

0.21I1

10

20

30
Bus Number

40

50

60

Fig. 9: Bus voltage profile before and after allocation

Table 1 displays the results obtained by GA. In Table 1 we have present rating and bus number of optimally allocation of SVC in IEEE57 Bus Test System (shown in FiglO).

206
[4] S.Gerbex, R.Cherkaoui and A.J. Germond, "Optimal Location of FACTS Devices in a Power System Using Genetic Algorithms ," in proceedings of the 13th power system computation conference,1999 pp .1252-1259. [5] C. A. Canizares, Modeling and Implementation of TCR and VSI Based FACTS Controllers, Internal report, ENEL and Politecnico di Milano, Milan, Italy (Oct. 1999). [6] C.A. Canizares, Massimo Pozzi, Sandro and Corsi, " SVC Modeling forVoltage and Angle Stability Studies". [7] C.Canizares, S.Corsi and M.Pozzi "Modeling and implementation of TCR and VSI based FACTS Controller;"AT-Unita controllo e regolazione No.99/595 December 1999. [8] Federico Milano, "Pricing System Security in Electricity Market Models with Inclusion of Voltage Stability Constraints, " PHD thesis, Genova, Italy, April 2003. [9] Holland.J.H, "Adaptation in Nature and Artifical Systems",the University of Michigan Press,1975. [10] Yang Hua Song and Malcom R. Irving. "Optimization Techniques for Electical Power System Part 2 Heuristic Optimization Methods".
[11] R. Seydel, "Practical Bifurcation and Stability Analysis: From Equilibrium to Chaos", Second Edition, Springer-Verlag, New York, 1994. [12] C. A. Canizares, Voltage Stability Assessment: Concepts, Practices and Tools," Tech. Rep., IEEE/PES Power System Stability Subcommittee, Final Document, Aug. 2002, available at http://www.power.uwaterloo.ca. [13] P. Kundur, "Power System Stability and Control", McGraw Hill, New York, 1 994. [14] A.C. Z. Souza, C.A. Canizares, and V. H. Quintana, "New Technique to Speed Up Voltage Collapse Computations Using Tangent Vector," IEEE Transactions on Power systems, vol. 12, no. 3, pp. 1380-1387, Aug. 1997. [15] C.A. Canizares and F.L. Alvarado, "Point of Collapse and Continuation Methods for Large AC/DC Systems," IEEE Transactions on Power Systems, vol. 8, no. 1, pp. 1{8, Feb. 1993. [16] V. Ajjarapu and C. Christy,"The continuation power Flow: a tool for steady state voltage stability analysis," IEEE Transactions on Power Systems, vol. 7, pp. 416{423, 1992. [17] Zeno T. Faur, "Effect of Facts Devices on Static Voltage Stability", Waterloo Ontario, Canada 1996.

Seyed hossein hosseinian received both the B.S. and M.S degree in Electrical Eng. Dept. Amirkabir University of technology, Iran, 1985,1988 respectively and PhD degree in Electrical Engineering Dept, university of Newcastle England, 1995. he is currently Assistant Professor of Electrical engineering Department in Amirkabir University of technology (AUT).

[18] C.A. Canizares, "Applications of Optimization to Voltage Collapse Analysis," Panel Session, Optimization Techniques in Voltage Collapse Analysis, IEEE/PES Summer Meeting, San Diego. [19] Naoto Yorino Shigemi Harada and Haozhong Cheng, "A Method for Approximate a Closest Loadability Limit Using Multiple Load Flow Solution", IEEE Trans on Power Systems, Vol.12 ,No. 1, February 1997.

VIII. BIOGRAPHIES
Sajad Najafi Ravadanegh received his B.S. degree in Electrical Engineering from University of Tabriz, Iran in 2001 and M.S. degree in Electrical Engineering from Amirkabir University of Technology (AUT),Iran in 2003.He is currently a Ph.D student in Dept. of Electrical Engineering at AUT,Iran. His research interests include power ststem stability and control, FACTS Devices, application of evolutionary algorithms in power system, nonlinear dynamic and chaos.
Mehrdad abedi received the B.S. degree in Electrical Eng. Dept. Tehran University, Iran, 1970, and the M.S from Electrical. Eng. Dept (power system division). Imperial College, University of London, England in 1973 and PhD degree Electrical. Eng. Dept (power system division).Newcastle University, England, 1977. he is currently full Professor of Electrical engineering Department in Amirkabir University of technology (aut) from 1997.