ARTICLE IN PRESS

Journal of the Franklin Institute 345 (2008) 166181 www.elsevier.com/locate/jfranklin

Optimal location and controller design of STATCOM for power system stability improvement using PSO
Sidhartha Panda, Narayana Prasad Padhy
Department of Electrical Engineering, Indian Institute of Technology, Roorkee, Uttarakhand-247667, India Received 14 February 2007; received in revised form 23 July 2007; accepted 1 August 2007

Abstract The optimal location of a static synchronous compensator (STATCOM) and its coordinated design with power system stabilizers (PSSs) for power system stability improvement are presented in this paper. First, the location of STATCOM to improve transient stability is formulated as an optimization problem and particle swarm optimization (PSO) is employed to search for its optimal location. Then, coordinated design problem of STATCOM-based controller with multiple PSS is formulated as an optimization problem and optimal controller parameters are obtained using PSO. A two-area test system is used to show the effectiveness of the proposed approach for determining the optimal location and controller parameters for power system stability improvement. The nonlinear simulation results show that optimally located STATCOM improves the transient stability and coordinated design of STATCOM-based controller and PSSs improve greatly the system damping. Finally, the coordinated design problem is extended to a four-machine two-area system and the results show that the inter-area and local modes of oscillations are well damped with the proposed PSO-optimized controllers. r 2007 The Franklin Institute. Published by Elsevier Ltd. All rights reserved.
Keywords: Static synchronous compensator; Transient stability; Optimal location; Particle swarm optimization; Power system stabilizer; Coordinated design; Power system stability

Corresponding author. Tel.: +91 9897563294.

E-mail address: panda_sidhartha@rediffmail.com (S. Panda). 0016-0032/$32.00 r 2007 The Franklin Institute. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jfranklin.2007.08.002

ARTICLE IN PRESS
S. Panda, N.P. Padhy / Journal of the Franklin Institute 345 (2008) 166181 167

1. Introduction Reactive power compensation is an important issue in electrical power systems and shunt exible AC transmission system (FACTS) devices play an important role in controlling the reactive power ow to the power network and, hence, the system voltage uctuations and stability. Static synchronous compensator (STATCOM) is member of FACTS family that is connected in shunt with the system. Even though the primary purpose of STATCOM is to support bus voltage by injecting (or absorbing) reactive power, it is also capable of improving the power system stability [1]. It has been proved that shunt FACTS devices give maximum benet from their stabilized voltage support when sited at the mid-point of the transmission line [2]. The rst swing stability of the system is greatly inuenced by choice of different models of the transmission line [3]. For long transmission lines, when the actual model of the line is considered, the results may deviate signicantly from those found for the simplied model. With pre-dened direction of real power ow, the shunt FACTS devices need to be placed slightly off-center towards the sending end for maximum benet from the transient stability point of view [4,5]. Application of GA to determine the optimal location of the shunt FACTS devices for transient stability improvement has also been reported in the literature [6]. Power system stabilizers (PSSs) are one of the most common controls used to damp out power system oscillations. When a STATCOM is present in a power system to support the bus voltage, a supplementary damping controller could be designed to modulate the STATCOM bus voltage in order to improve damping of system oscillations [7]. But, the interaction among PSSs and STATCOM-based controller may enhance or degrade the damping of certain modes of rotors oscillating modes. To improve overall system performance, many researches were made on the coordination between PSSs and FACTS power oscillation damping controllers [811]. Although the local control signals are easy to get, they are not as highly controllable and observable as wide area signals for the inter-area oscillation modes. Due to restriction of local measurements, these controllers based on local signals tend to be difcult to offer satisfactory performance under various system operating conditions. With the rapid advancement in wide area measurement systems technology, fast communication networks and powerful information technology, the widely dispersed signals of power systems can be centralized, processed and distributed even in real time, which makes the wide area signal a good alternative for control input [12]. A number of conventional techniques have been reported in the literature pertaining to design problems of conventional PSSs: the eigenvalue assignment, mathematical programming, gradient procedure for optimization and also the modern control theory. Unfortunately, the conventional techniques are time consuming as they are iterative and require heavy computation burden and slow convergence. In addition, the search process is susceptible to be trapped in local minima and the solution obtained may not be optimal [13]. The evolutionary methods constitute an approach to search for the optimum solutions via some form of directed random search process. A relevant characteristic of the evolutionary methods is that they search for solutions without previous problem knowledge. Recently, particle swarm optimization (PSO) appeared as a promising evolutionary technique for handling the optimization problems. PSO is a population-based stochastic optimization technique, inspired by social behavior of bird ocking or sh schooling [14]. This paper proposes to use PSO technique to determine optimal location of STATCOM for transient stability improvement.

ARTICLE IN PRESS
168 S. Panda, N.P. Padhy / Journal of the Franklin Institute 345 (2008) 166181

For damping the power system oscillations and for improving the interactions between PSSs and STATCOM-based controller, PSO-based optimal tuning approach is employed to coordinately design the proposed damping controllers. The difference of speed deviations is taken as the input signal for the proposed STATCOM-based controller. The reminder of the paper is organized as follows. A two-area system with a STATCOM is described in Section 2. In Section 3, a brief overview of PSO is provided. Optimal location of STATCOM and its optimal controller parameters are formulated as optimization problems in Section 4. The computer simulation results for system under study are presented and discussed in Section 5. 2. Two-area system with STATCOM Consider a two-area system (areas 1 and 2), connected by a long transmission line with a STATCOM as shown in Fig. 1. The direction of real power ow (PL) is assumed to be from area-1 to -2. If the rating of the STATCOM is large enough to supply the reactive power required to maintain constant voltage magnitude at the point of connection, it effectively divides the transmission line into two sections (sections 1 and 2). In Fig. 1, L is the distance from the sending end, at which the STATCOM is located. The power ows at the sending end and receiving end for a long transmission line with distributed parameters can be written as [15] PS K 1 cos yB yA K 2 cos yB d, PR K 2 cos yB d K 3 cos yB yA , where K 1 AV 2 S =B; and A jAjyA ; V R jV R j0; B jBjyB; V S jV S jd. K 2 AV S V R =B; K 3 AV R =B (1) (2)

It is clear from Eq. (2) that the receiving-end real power PR reaches the maximum value when the angle d becomes yB. However, the sending-end real power PS of Eq. (1) becomes maximum at d 180yB.
Area-1 VS section-1 L PL Load-1 VST VSC VDC
Fig. 1. Two-area system with STATCOM.

VM section-2

VR

Area-2

Sending-end

iST Load-2

Receiving-end

ARTICLE IN PRESS
S. Panda, N.P. Padhy / Journal of the Franklin Institute 345 (2008) 166181 169

15 b c SEP & REP [pu] 10 a

SEP REP

0 0

3 50

2 150 200

100 Power angle [deg]

Fig. 2. Sending-end and receiving-end power-angle characteristics for actual line model.

The power-angle characteristic of the line using the actual line model is shown in Fig. 2 (without the STATCOM). Fig. 2 is drawn using Eqs. (1) and (2) by varying the power angle d, from 01 to 1801. It is noticed that when the angle d increases from zero, both the sending-end power (SEP) and receiving-end powers (REP) are increased. As the line resistances are not neglected, the angle yB is slightly less than 901. Hence, as d increases, rst the angle d becomes equal to yB, at which point the REP becomes maximum (point a in Fig. 2). With further increase in d, it becomes equal to 180yB, at which point SEP becomes maximum (point b in Fig. 2). So, REP rst reaches the maximum value at an angle d1 followed by the maximum SEP at an angle d2. Now let us consider that the STATCOM is placed at the mid-point of the line as shown in Fig. 1, and its rating is large enough to supply the reactive power required to maintain a constant voltage magnitude at mid-point. So the transmission line is divided into two equal sections and the SEP and REP characteristic of two equal sections are identical. Each section can be represented by the same power-angle characteristic curve of Fig. 2. But, from Fig. 2, it is clear that maximum value of SEP is more than the maximum value of REP for a line section. Hence, for the two equal sections with identical power-angle characteristic, maximum SEP of section-2 (SEP-2) is more than maximum REP of section-1 (REP-1). If section-1 delivers the maximum power at its receiving-end (point a, in Fig. 2) the corresponding SEP of section-2 can be represented by the same power level (point c, in Fig. 2). So, even though the section-2 is capable of carrying more power, it carries that much power which it receives at its receiving-end. Hence the maximum power transfer capability of the system is limited by the maximum REP-1. The total transmission angle at the maximum power point is d d1+d3 and if the power angle is further increased the power transfer through the line decreases because of the decrease in maximum REP-1 and section-1 operates in unstable region. When maximum REP-1 is greater than maximum SEP-2, section-2 operates in the unstable region at higher power angles and Pd characteristics follows REP curve. However, when maximum REP-1 is less than maximum SEP-2, section-1 operates in the unstable region at higher power angles and Pd characteristics follows SEP curve. The above unusual change in the pattern of the Pd curve which signicantly affects the stability of the system can be utilized to improve the system stability. As for a given initial operating condition and fault clearing time the area

ARTICLE IN PRESS
170 S. Panda, N.P. Padhy / Journal of the Franklin Institute 345 (2008) 166181

between the Pd curve and the initial operating power line is a measure of the decelerating area. In equal-area criterion method this area is used to determine the stability of the system. This area can be increased by shifting the Pd curve towards the left by shifting the location of the STATCOM towards the sending-end, and the stability can be further improved. The SEP and REP power-angle characteristics depend upon the length of the line section. For lower values of L (i.e., as the location of the STATCOM is moved towards the sending end), the length of line section-1 decreases and the length of line section-2 increases. Hence, the maximum REP-1 increases while the maximum SEP-2 decreases. Thus the point a in Fig. 2 moves upwards and point b goes downwards. Maximum REP-1 will be equal to maximum SEP-2 at some off-center location of STATCOM. In other words, as the location of STATCOM is moved towards the sending-end, at some value of L, the points a and b become equal and are at the same power level. The location of STATCOM at which the points a and b are equal is the optimal location from transient stability improvement point of view. In the present paper, PSO technique is applied to determine the optimal location of STATCOM. 3. Overview of particle swarm optimization PSO method is a member of wide category of swarm intelligence methods for solving the optimization problems. It is a population-based search algorithm where each individual is referred to as particle and represents a candidate solution. Each particle in PSO ies through the search space with an adaptable velocity that is dynamically modied according to its own ying experience and also to the ying experience of the other particles. In PSO, each particle strive to improve itself by imitating traits from their successful peers. Further, each particle has a memory and hence it is capable of remembering the best position in the search space ever visited by it. The position corresponding to the best tness is known as pbest and the overall best out of all the particles in the population is called gbest [16,17]. The modied velocity and position of each particle can be calculated using the current velocity and the distance from the pbestj,g to gbestg as shown in the following formulas [18]:
1 t t t vj;tg wv j ;g c1 r1 pbestj ;g xj ;g c2 r2 gbestg xj ;g , 1 t t1 x xj;tg j ;g vj ;g .

(3) (4)

with j 1; 2; . . . ; n and j 1; 2; . . . ; m, where n is the number of particles in a group; m the t number of members in a particle; t the number of iterations (generations); v j ;g thevelocity
t max of particle j at iteration t, V min g pvj ;g pV g ; w the inertia weight factor; c1 and c2 are the cognitive and social acceleration factors, respectively; r1 and r2 are the random numbers t uniformly distributed in the range (0, 1); x j ;g is the current position of particle j at iteration t; pbestj the pbest of particle j; gbest the gbest of the group. The j-th particle in the swarm is represented by a g-dimensional vector xj xj;1 ; xj;2 ; . . . ; xj ;g and its rate of position change (velocity) is denoted by another g-dimensional vector vj vj ;1 ; vj ;2 ; . . . ; vj ;g . The best previous position of the j-th particle vector pbestj pbestj ;1 ; pbestj;2 ; . . . ; pbestj;g . The index of best particle among all the particles in the group is represented by gbestg. In PSO, each particle moves in the search space with a velocity according to its own previous best solution and its groups previous

ARTICLE IN PRESS
S. Panda, N.P. Padhy / Journal of the Franklin Institute 345 (2008) 166181 171

best solution. The velocity update in a PSO consists of three parts: momentum, cognitive and social parts. The balance among these parts determines the performance of a PSO algorithm. The parameters c1 and c2 determine the relative pull of pbest and gbest and the parameters r1 and r2 help in stochastically varying these pulls. In the above equations, superscripts denote the iteration number. 4. Problem formulation 4.1. Optimal location of STATCOM As explained in Section 2, the power transfer capability and hence the transient stability of the system can be improved by locating STATCOM slightly off-center towards the sending-end instead of the mid-point. For a given initial operating conditions, there will be an optimal location of STATCOM where the maximum sending-end power of section-1 is equal to the maximum receiving-end power of section-2. From transient stability point of view, the value of maximum sending-end power is important as the power that can be drawn from generator terminals immediately after fault clearance inuences the transient stability. Therefore, the objective function to maintain transient stability can be written in the following form: Objective function J jmaximumDd1 Dd1 j, (5)

where Dd1 and Dd2 are the rotor angle deviation following a disturbance of generators in areas 1 and 2, respectively, and |maximum(Dd1Dd2)| is the absolute value of maximum rotor angle deviation difference of two areas. If |maximum(Dd1Dd2)|o1801 the system is stable. For objective function calculation, the time-domain simulation of the nonlinear system model is carried out for the simulation period. PSO is employed to search for the location of STATCOM where the value of objective function is minimum. The problem constraints are the location bounds. Therefore, the design problem can be formulated as the following optimization problem: minimize J subject to Lmin pLpLmax , (7) (6)

where L is the length of line section from the sending-end to the location of STATCOM. The following steps are followed to search for the optimal location of STATCOM to improve transient stability: Step-1: Initially set the fault clearing time TFC to a high value so that the system is unstable at all locations of STATCOM. Step-2: Employ PSO to minimize the objective function J Step-3: Check for stability of the system. Step-4: If the system is unstable, decrease TFC by a small step and repeat from Step-2. Stop if the system is stable. The system will be stable at TFC TFCF only if the STATCOM is placed at optimal location obtained by the above method and for TFC4TFCF the system becomes unstable at all locations.

ARTICLE IN PRESS
172 S. Panda, N.P. Padhy / Journal of the Franklin Institute 345 (2008) 166181

4.2. Coordinated design of STATCOM-based controller and PSS The commonly used leadlag structure is chosen in this study as STATCOM-based controller as shown in Fig. 3. The structure consists of: a gain block, a signal washout block and two-stage phase-compensation block. The phase-compensation block provides the appropriate phase-lead characteristics to compensate for the phase lag between input and the output signals. The signal washout block serves as a high-pass lter which allows signals associated with oscillations in input signal to pass unchanged. Without it steady changes in input would modify the output. From the viewpoint of the washout function the value of washout time constant is not critical and may be in the range 120 s [19]. In this structure, the washout time constants TWS and the time constants T2s, T4s are usually prespecied. In the present study, TW 10 s and T2s T4s 0.3 s are used. The controller gain KS and the time constants T1s and T3s are to be determined. The generic PSS block of the SimPowerSystem (SPS) toolbox is used to add damping to the rotor oscillations of the synchronous machine by controlling its excitation. The output signal of the PSS is used as an additional input (VS) to the excitation system block. The PSS input signal can be either the machine speed deviation or acceleration power. The PSS model consists of a low-pass lter, a general gain, a washout high-pass lter, a phasecompensation system, and an output limiter as shown in Fig. 4. The general gain KP determines the amount of damping produced by the stabilizer. The washout high-pass lter eliminates low frequencies that are present in the input signal and allows the PSS to respond only to changes in the input. The phase-compensation system is used to compensate the phase lag between the excitation voltage and the electrical torque of the synchronous machine. In this structure, a washout time constant TWP 3 s is used. The time constants T1P and T2P are to be determined. It is worth mentioning that the PSS and STATCOM-based controllers are designed to minimize the power system oscillations after a large disturbance so as to improve the power system stability. In the present study, an integral time absolute error of the speed
V ST KS Input Gain block sTWS 1 + sTWS Washout block 1 + sT1s 1 + sT2S 1 + sT3S 1 + sT4S
min V ST max

VST +

VST Output

Two-stage lead-lag block

VST_Ref

Fig. 3. Structure of the STATCOM-based controller.

VS Input Sensor KP Gain block sTWP 1 + sTWP Washout block 1 + sT1P 1 + sT2P Lead-lag block

max

VS Output
min VS

Fig. 4. Structure of the generic power system stabilizer.

ARTICLE IN PRESS
S. Panda, N.P. Padhy / Journal of the Franklin Institute 345 (2008) 166181 173

deviations is taken as the objective function expressed as follows: Z ttsim jDojt dt, J
t0

(8)

where Do denotes the speed deviation for a set of controller parameters, and tsim is the time range of the simulation. For objective function calculation, the time-domain simulation of the nonlinear power system model is carried out for the simulation period. It is aimed to minimize this objective function in order to improve the system response in terms of the settling time and overshoots. In this study, it is aimed to minimize the proposed objective function J. The problem constraints are the PSS and STATCOM controller parameter bounds. Therefore, the design problem can be formulated as the following optimization problem: Minimize J subject to
max K min S pK S pK S , max T min 1S pT 1S pT 1S , max T min 3S pT 3S pT 3S , max K min P1 p K P1 p K P 1 , max T min 1P1 pT 1P1 pT 1P1 , max T min 2P1 pT 2P1 pT 2P1 , max K min P2 p K P2 p K P 2 , max T min 1P2 pT 1P2 pT 1P2 , max T min 2P2 pT 2P2 pT 2P2 .

(9)

10

The proposed approach employs PSO technique to solve this optimization problem and searches for the optimal set of PSS and STATCOM-based controller parameters. 5. Results and discussions The SPSs toolbox is used for all simulations and STATCOM-based controller design. SPS is a MATLAB-based modern design tool that allows scientists and engineers to rapidly and easily build models to simulate power systems using Simulink environment. The SPSs main library, powerlib, contains models of typical power equipment such as machines, governors, excitation systems, transformers, lines and FACTS devices. The library also contains the powergui block that opens a graphical user interface for the steady-state analysis of electrical circuits. The load ow and machine initialization option of the powergui block performs the load ow and the machines initialization [20]. In order to determine the optimal location of STATCOM and design the STATCOMbased controller, the MATLAB/SIMULINK model of the test system depicted in Fig. 1 is developed using SPS blockset. The system consists of two hydraulic generating units, one of 1400 MVA in one area and 700 MVA in the other. The generators are represented by a sixth-order model and both are equipped with hydraulic turbine and governor (HTG)

ARTICLE IN PRESS
174 S. Panda, N.P. Padhy / Journal of the Franklin Institute 345 (2008) 166181

and excitation system. The HTG represents a nonlinear hydraulic turbine model, a PID governor system, and a servomotor. The excitation system consists of a voltage regulator and DC exciter, without the exciters saturation function. All the relevant parameters are given in Appendix. The generators with output voltages of 13.8 kV are connected by a long 500 km transmission line through three-phase step-up transformers. The output voltage of transformer is 500 kV. The loads in each area are so chosen that the real power ow on the transmission line is always from area-1 to -2. The STATCOM used for this model is a phasor model with a rating of 7200 MVA. The pre-fault reference voltage is set to 1 pu for STATCOM (please refer to Appendix for other parameters). 5.1. Application of PSO to determine optimal location Initial power outputs of the generators chosen are P1 0.75 pu and P2 0.4 pu. The loads at the area-1 are 160 MW and 200 MVAR and those at area-2 are 1340 MW and 500 MVAR. The reference voltage of STATCOM is set 1.0 pu. The pre-fault SEP and REP are 918 and 886 MW, respectively. A three-phase fault is applied at the sending-end bus at time t 0.1 s. The original system is restored upon the clearance of the fault. Also to improve transient stability the reference voltage is increased to 1.1 pu immediately after the fault clearance. The optimal location of STATCOM is obtained employing PSO using the steps given in Section 4.1. For the optimization of objective function given in Eq. (6), routines from PSO toolbox are used. While applying PSO technique, a number of parameters are required to be specied. An appropriate choice of the parameters affects the speed of convergence of the algorithm. Table 1 shows the parameters used in the present study for the PSO algorithm. The optimal location of STATCOM is found to be at L 212 km from the sending end and the corresponding highest critical fault clearing time TFCF 0.079 s by the above PSOoptimization approach. This information about the optimal location is useful during the planning stage so that the substation can be located at the optimal location (or nearest to the optimal location as far as possible) to install the STATCOM for maximum benet in terms of both transient stability and power handling capacity. To verify the obtained result, the above contingency (three-phase fault at sending-end at t 0.1 s and cleared at t 0.179 s) is simulated for different locations of STATCOM. Fig. 5 shows the variation of rotor angle difference (d1d2) for different locations of STATCOM which conrms that the system is stable for the given initial operating conditions and fault clearing time only if the STATCOM is placed at the optimal location (L 212 km) and unstable at all other locations.

Table 1 Parameters used for PSO algorithm PSO parameters Swarm size No. of generations c1, c2 wstart, wend Value/type 20 100 2.0, 2.0 0.9, 0.4

ARTICLE IN PRESS
S. Panda, N.P. Padhy / Journal of the Franklin Institute 345 (2008) 166181 175

Fig. 5. Variation of rotor angle difference for different locations of STATCOM (TFC 0.079 s and SEP 918 MW).

5.2. Application of PSO to determine optimal controller parameters Owing to the recent advances in optical ber communication and global positioning system, wide area measurement system can realize phasor measurement synchronously and deliver it to the control center even in real time, which makes the wide area signal a good alternative for control input. In view of the above, the difference of speed deviations of generators in the two areas (o1o2) is chosen as the control input of STATCOM-based controller in this paper. Also, accelerating powers of the individual generators are chosen as the input signal for the two PSSs. The parameters of the PSS and the STATCOM-based controller are optimally tuned using PSO technique. The objective function given in Eq. (8) is evaluated for each individual by simulating the example power system, considering a severe three-phase fault disturbance. The convergence of objective function J with the number of generations is shown in Fig. 6. The obtained parameters of STATCOM-based controller and PSS are shown in Table 2. To evaluate the capability of the PSO-optimized PSS and STATCOM-based controller, time-domain simulation is performed on the example power system under various severe and small disturbances. The system response with controllers without the optimized parameters is shown with dotted line and the response with controllers with optimized parameters is shown with solid lines. Note that, in controller without the optimized parameters case, both STATCOM (with Vref 1.0 pu before fault and 1.1 pu after fault clearance) and PSS are present in the system, but the STATCOM-based controller is not considered and the parameters of the PSSs are not optimized. In order to show the effectiveness of the PSO-optimized controllers, the same contingency is applied and the variation of (d1d2) against time is shown in Fig. 7. It is clear from Fig. 7 that the power system oscillations are quickly damped out with application of proposed PSO-optimized controllers. The effectiveness of the proposed controllers to variation in line power ow is also examined. By changing the loads in each area, the SEP is increased by 100 MW (from 918 to 1018 MW) and a three-phase fault is applied at the sending-end bus at t 0 s and

ARTICLE IN PRESS
176 S. Panda, N.P. Padhy / Journal of the Franklin Institute 345 (2008) 166181

1.522 Convergence of J 1.52 1.518 1.516 1.514 1.512 1.51 0 10 20 30 40 50 60 Generation 70 80 90 100

Fig. 6. Convergence of objective function.

Table 2 PSO-optimized parameters of STATCOM-based controller and PSSs STATCOM-base controller KS 103.6343 T1S 0.5703 T3S 0.0493 PSS-1 parameters KP1 0.9788 T1P1 0.0837 T2P1 0.7683 PSS-2 parameters KP2 0.3264 T1P2 0.0893 T2P2 0.8920

Fig. 7. Variation of rotor angle difference without and with PSO-optimized controllers (TFC 0.079 s and SEP 918 MW).

cleared after 0.077 s. Fig. 8 shows the variation of the rotor angle difference of the two machines for controllers without the optimized parameters and the controllers with the optimized parameters.

ARTICLE IN PRESS
S. Panda, N.P. Padhy / Journal of the Franklin Institute 345 (2008) 166181 177

Fig. 8. Variation of rotor angle difference without and with PSO-optimized controllers at higher sending-end power (TFC 0.075 s and SEP 1018 MW).

Fig. 9. Variation of rotor speed difference under small disturbance.

To show the effectiveness of the PSO-optimized controllers under small disturbance, the load at receiving end is disconnected for 50 ms (this simulates a small disturbance). The variation of speed deviation difference of the two machines and STATCOM bus voltage (VST) are shown in Figs. 9 and 10. From the gure it is clear that the performance of the power system has been improved signicantly by modulating the STATCOM bus voltage following the disturbance and the damping has been improved considerably. 5.3. Design problem for Kundurs four-machine, two-area system The design problem is further extended to Kundurs four-machine, two-area system shown in Fig. 11 [19]. Speed deviations of generators G1 and G4 are selected as the input signal of the STATCOM-based controller. Accelerating power of the individual generators are chosen as the input signals for all four PSSs.

ARTICLE IN PRESS
178 S. Panda, N.P. Padhy / Journal of the Franklin Institute 345 (2008) 166181

Fig. 10. Variation of STATCOM reference voltage signal under small disturbance.

1 G1

5 25 km

6 10 km L-1

8 110 km 110 km

9 10 km L-2

10 25 km

11

3 G3

Area 1

2 G2

STATCOM G4

Area 2

Fig. 11. Four-machine two-area system with STATCOM.

An integral time absolute error of the speed signals corresponding to the local and interarea modes of oscillations is taken as the objective function. The objective function is expressed as Z ttsim X  X J DoL DoI t dt, (11)
t0

where DoL and DoI are the speed deviations of local and inter-area modes of oscillations, respectively, and tsim is the time range of the simulation. The parameters of the PSSs and the STATCOM-based controller are optimally tuned using PSO technique as explained above. The convergence of objective function J with the number of generations is shown in Fig. 12. The obtained parameters of STATCOM-based controller and PSS are shown in Table 3. To show the effectiveness of the proposed PSO-optimized controllers based on wide area signal under extreme conditions, time-domain simulation is performed on the system with a three-phase fault applied at the sending end of the circuit between buses 7 and 8 (near bus 7) that is cleared 100 ms later. Figs. 13 and 14 show the variations of the inter-area and local mode of oscillation, against time, respectively. From these gures, it can be seen that both inter-area and local modes of oscillations are highly

ARTICLE IN PRESS
S. Panda, N.P. Padhy / Journal of the Franklin Institute 345 (2008) 166181 179

6.5 6 Convergence of J 5.5 5 4.5 4 3.5 0 20 40 60 80 100 120 Generation 140 160 180 200

Fig. 12. Convergence of objective function. Table 3 PSO-optimized parameters of STATCOM-based controller and PSSs for four-machine two-area system Controller/parameters Gain Time constant STATCOM 232.323 0.2892 0.2641 PSS-1 3.5952 0.0801 1.7881 PSS-2 4.6062 0.0396 0.6691 PSS-3 4.3768 0.0758 1.0796 PSS-4 4.2875 0.0699 1.2285

Fig. 13. Inter-area mode of oscillation for 100 ms three-phase fault near bus 7.

oscillatory for the case of controllers without optimized parameters. For the case of controllers with PSO-optimized parameters, these modal oscillations are quickly damped out.

ARTICLE IN PRESS
180 S. Panda, N.P. Padhy / Journal of the Franklin Institute 345 (2008) 166181

Fig. 14. Local mode of oscillation for 100 ms three-phase fault near bus 7.

6. Conclusions In this paper, PSO technique is applied to determine the optimal location and controller parameters of STATCOM. First, a systematic procedure to determine the optimal location of STATCOM for transient stability improvement following a severe disturbance is proposed. The proposed algorithm is applied to nd the optimal location of STATCOM in a two-area system employing PSO. Further, a STATCOM-based damping controller is proposed and the parameter of the proposed controller and PSSs are coordinately determined using PSO. The obtained nonlinear simulation results show that optimally located STATCOM extends the critical fault clearing time and hence improve transient stability of the system. Also, coordinated design of STATCOM-based controller and PSSs using PSO technique, improve greatly the system stability by damping out the power system oscillations quickly, under severe and small disturbance conditions. Finally, the coordinated design problem is extended to a four-machine two-area system. The timedomain simulations have shown that the inter-area and local modes of oscillations are well damped with the PSO-optimized controllers. Appendix The data for various components used in the MATLAB simulation (All data are in pu unless specied otherwise; the notations used are as in SimPowerSystem toolbox.): Generator parameters: M1 1400 MVA, M2 700 MVA, V 13.8 KV, f 60 Hz, Xd 1.305, Xd1 0.296, Xd00 0.255, Xq 0.474, Xq00 0.243, X1 0.18. Transformer parameters: T1 1400 MVA, T2 700 MVA, 13.8/500 KV, R2 0.002, L2 0.12, Rm 500 O, Xm 500 O. Transmission line parameters per km: R1 0.1755 O, R0 0.2758 O, L1 0.8737 mH, L0 3.22 mH, C1 13.33 nF, C0 8.297 nF.

ARTICLE IN PRESS
S. Panda, N.P. Padhy / Journal of the Franklin Institute 345 (2008) 166181 181

STATCOM parameters: 500 KV, 7200 MVAR, R 0.071, L 0.22, VDC 40 KV, CDC 3757mF, Vref 1.0, KP 50, Ki 1000. min STATCOM controller: DVmax ST 1.1 pu, DVST 0.9 pu, T2S T4S 0.3 s, TWS 10 s. max PSSs: sensor time constant 0.015 s, VS 0.15 pu, Vmin 0.15 pu. S References
[1] N.G. Hingorani, L. Gyugyi, Understanding FACTS: Concepts and Technology of Flexible AC Transmission Systems, IEEE Press, New York, 2000. [2] B.T. Ooi, M. Kazerani, R. Marceau, Z. Wolanski, F.D. Galiana, D. McGills, G. Joos, Mid-point siting of FACTS devices in transmission lines, IEEE Trans. Power Del. 12 (1997) 17171722. [3] M.H. Haque, Effects of exact line model and shunt FACTS devices on rst swing stability limit, Int. J. Power Energy Syst. 25 (2005) 121127. [4] S. Panda, R.N. Patel, Improving power system transient stability with an off-centre location of shunt FACTS devices, J. Electr. Eng. 57 (9) (2006) 365368. [5] M.H. Haque, Optimal location of shunt FACTS devices in long transmission lines, IEE Proc. Gen. Trans. Distrib. 147 (2000) 218222. [6] S. Panda, R.N. Patel, Optimal location of shunt FACTS controllers for transient stability improvement employing genetic algorithm, Electr. Power Comp. Syst. 35 (2007) 189203. [7] H.F. Wang, PhillipsHeffron model of power systems installed with STATCOM and applications, IEE Proc. Gen. Trans. Distrib. 146 (1999) 521527. [8] J.M. Ramirez, I. Castillo, PSS and FDS simultaneous tuning, Electr. Power Syst. Res. 68 (2004) 3340. [9] Q.J. Liu, Y.Z. Sun, T.L. Shen, Y.H. Song, Adaptive nonlinear coordinated excitation and STATCOM controller based on Hamiltonian structure for multimachine-power-system stability enhancement, IEE Proc. Gen. Trans. Distrib. 150 (2003) 285294. [10] X. Lei, X. Li, D. Povh, A nonlinear control for coordinating TCSC and generator excitation to enhance the transient stability of long transmission, Electr. Power Syst. Res. 59 (2001) 103109. [11] L.J. Cai, I. Erlich, Simultaneous coordinated tuning of PSS and FACTS damping controller in a large power system, IEEE Trans. Power Syst. 20 (2005) 294300. [12] Y. Chang, Z.Xu, A novel SVC supplementary controller based on wide area signals, Electr. Power Syst. Res., (2006), doi:10.1016/j.epsr.2006.11.002. [13] Y.L. Abdel-Magid, M.A. Abido, Robust coordinated design of excitation and TCSC-based stabilizers using genetic algorithms, Int. J. Electr. Power Energy Syst. 69 (2004) 129141. [14] J. Kennedy, R.C. Eberhart, Particle swarm optimization, in: Proceedings of the IEEE International. Conference on Neural Networks, vol. 4, Piscataway, NJ, IEEE Service Center, 1995, pp. 19421948. [15] H. Saadat, Power System Analysis, Tata McGraw-Hill, New Delhi, 2002. [16] J. Kennedy, R. Eberhart, Swarm Intelligence, Academic press, San Diego, CA, 2001. [17] PSO Tutorial, /http://www.swarmintelligence.org/tutorials.phpS. [18] Z.L. Gaing, A particle swarm optimization approach for optimum design of PID controller in AVR system, IEEE Trans. Energy Convers. 9 (2004) 384391. [19] P. Kundur, Power System Stability and Control, McGraw-Hill, New York, 1994. [20] SimPowerSystems 4.3 Users Guide, Available: /http://www.mathworks.com/products/simpower/S.