Electric Power Systems Research 78 (2008) 17261735

Contents lists available at ScienceDirect

Electric Power Systems Research

journal homepage: www.elsevier.com/locate/epsr

A self-tuning fuzzy PI controller for TCSC to improve power system stability

Salman Hameed, Biswarup Das , Vinay Pant
Department of Electrical Engineering, Indian Institute of Technology, Roorkee, Roorkee 247667, Uttarakhand, India

a r t i c l e

i n f o

a b s t r a c t
In this paper, a self-tuning fuzzy PI controller (STFPIC) is proposed for thyristor-controlled series capacitor (TCSC) to improve power system dynamic performance. In a STFPIC controller, the output-scaling factor is adjusted on-line by an updating factor (). The value of is determined from a fuzzy rule-base dened on error (e) and change of error ( e) of the controlled variable. The proposed self-tuning controller is designed using a very simple control rule-base and the most natural and unbiased membership functions (MFs) (symmetric triangles with equal base and 50% overlap with neighboring MFs). The comparative performances of the proposed STFPIC and the standard fuzzy PI controller (FPIC) have been investigated on two multi-machine power systems (namely, 4 machine, 2 area system and 10 machine 39 bus system) through detailed non-linear simulation studies using MATLAB/SIMULINK. From the simulation studies it has been found out that for damping oscillations, the performance of the proposed STFPIC is better than that obtained by the standard FPIC. Moreover, the proposed STFPIC as well as the FPIC have been found to be quite effective in damping oscillations over a wide range of operating conditions and are quite effective in enhancing the power carrying capability of the power system signicantly. 2008 Elsevier B.V. All rights reserved.

Article history: Received 7 January 2008 Received in revised form 5 March 2008 Accepted 6 March 2008 Available online 24 April 2008 Keywords: Thyristor-controlled series capacitor Self-tuning fuzzy controller Power system stability

1. Introduction For economic and ecological reasons, the building of new transmission lines and expansion of existing transmission systems are becoming more and more difcult. In this new situation, it is necessary to utilize the existing power transmission system at its maximum capacity to meet increasing demand of electrical energy. However, the power transfer capability of an interregional AC transmission system is usually limited by the stability problems. As a result, power utilities are now placing more emphasis on improving the stability limits of the existing systems to increase the utilization of existing transmission facilities. In this context, it is nowadays well recognized that by applying the exible AC transmission system (FACTS) controllers, the stability limits can be enhanced signicantly [1,2]. Among various FACTS controllers, thyristor-controlled series capacitor (TCSC) is one of the most promising FACTS devices having a few practical installations around the world [3,4] and has attracted a lot of attention for designing an effective control law to enhance the system stability. The various control schemes reported in the literature for TCSC can be classied into two broad categories: (a) linearised, eigenvalue analysis based control system and (b) intelligent technique-based control scheme. Although the effectiveness of the eigenvalue analysis based control scheme has been proven in several publications, as pointed out in [5], it is neither
Corresponding author. E-mail addresses: biswafee@yahoo.com, biswafee@iitr.ernet.in (B. Das). 0378-7796/$ see front matter 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.epsr.2008.03.005

simple to develop the linearised system model nor is absolutely necessary for developing a FACTS damping controller. As a result, different intelligent technique-based controllers for TCSC have been suggested in the literature. Fang et al. [5] have proposed an OTEF descent strategy for designing fuzzy TCSC damping controller. In this work, the TCSC controller actually consists of two TCSC fuzzy controllers and the efcacy of the developed controller has been tested on a four-generator, two area interconnected power system. In Ref. [6], the authors have presented a TS fuzzy model scheme for TCSC which has been tested on a single machine innite bus (SMIB) system. Dash et al. have suggested a hybrid fuzzy controller and a non-linear TS fuzzy controller for TCSC in [7] and [8], respectively. Both these schemes have been tested on a three machine, six bus system with two TCSCs installed in the study system. In Ref. [9], the authors have proposed a new design technique, namely F-HGAPSO, to design the fuzzy controller. The effectiveness of their proposed controller has been tested on a SMIB system. Laiq Khan and Lo [10] have presented a hybrid micro-GA based fuzzy controller for TCSC. The performance of the proposed TCSC controller has been tested on the three machine, nine bus system. However, in this work, both TCSC and UPFC were considered in the study system. In Ref. [11], the authors have proposed a combination of a fuzzy controller and a conventional PI controller for TCSC and the validity of this strategy has been tested on a two area four-machine power system. From the above discussion it is observed that the different fuzzy control strategies proposed in the literature have been tested on relatively small test systems. This paper aims to extend the work on

S. Hameed et al. / Electric Power Systems Research 78 (2008) 17261735

1727

Fig. 1. Block diagram of the PI-type FLC (FPIC).

Fig. 2. MFs for e,

e and

u. N: negative; P: positive; ZE: zero; B: big; M: medium; S: small.

application of intelligent control technique for TCSC control design further. Specically, in this paper, a new self-tuning fuzzy PI control for TCSC is proposed to enhance the power system stability. Further, the effectiveness of the developed TCSC controller has been tested on a relatively large 10 machine 39 bus system (which has not been used in earlier publications [511]). This paper is organized as follows. Section 2 describes the proposed fuzzy logic controller for TCSC. Section 3 presents the main results of this work. Finally, Section 4 discusses the conclusions of this work.

Finally, the actual value of the controller output (u) is computed by u(k) = u(k 1) + u(k) (2)

The relationships between the SFs (Ge , G e and Gu ) and the input and output variables of the FPIC are as follows: eN = Ge e eN = G u = Gu
e

e uN

2. Fuzzy PI controller The block diagram of the fuzzy PI controller (FPIC) is shown in Fig. 1 [12,13]. In this gure, e(k) is the error at the kth sample and it can be written as e(k) = ysp y(k) where, y(k) is the actual system output and ysp the set-point or desired system output at kth sample, respectively. The change in error is dened as e(k) = e(k) e(k 1) (1)

The quantities e and e are converted to normalized quantities eN and eN , respectively by using the scaling factors (SFs) Ge and G e . These normalized quantities eN and eN are crisp in nature and therefore need to be rst converted to their corresponding fuzzy variables. After fuzzication, the fuzzied inputs are given to the fuzzy inference mechanism which, depending on the given fuzzy rule base, gives the normalized incremental change in control output ( uN ). The output uN is converted into actual incremental change in control output ( u) by using the scaling factor Gu . For implementing the fuzzy inference engine, the min operator for connecting multiple antecedents in a rule, the min implication operator, and the max aggregation operator have been used. Actually, the output uN from the inference mechanism is fuzzy in nature, hence, to determine the crisp output, these fuzzy outputs need to be defuzzied. The centroid defuzzication scheme has been used here for obtaining the output u as shown in Fig. 1.

Here Ge , G e and Gu are the SFs for e, e and u, respectively and eN , eN and uN are normalized quantities. The SFs are the main parameters used for tuning any fuzzy logic controller (FLC) because variation of the SFs changes the normalized universe of discourse of input and output variables and their corresponding membership functions. Generally, selection of suitable values for Ge , G e and Gu are made based on the knowledge about the process to be controlled and sometimes through trial and error to achieve the best possible control performance. This is so because, unlike conventional non-fuzzy controllers, there is no well-dened method for selecting appropriate values of SFs for FLC. However, if required, it is possible to tune these parameters to achieve a given control objective using some optimization techniques. In this work, the appropriate values for Ge , G e and Gu have been determined
Table 1 Rule base for e/e NB NM NS ZE PS PM PB NB NB NB NB NB NM NS ZE

u NM NB NM NM NM NS ZE PS NS NB NM NS NS ZE PS PS ZE NM NM NS ZE PS PM PM PS NS NS ZE PS PS PM PB PM NS ZE PS PM PM PM PB PB ZE PS PM PB PB PB PB

1728

S. Hameed et al. / Electric Power Systems Research 78 (2008) 17261735 Table 2 Ranges for the different scaling factors Parameters Minimum range Maximum range Ge 0 0.5 G 0 1
e

Gu 0 1

Table 3 Parameters used in genetic algorithm Parameter Maximum generations Population size Mutation rate Crossover operator Value/Type 50 25 0.1 Scattered

Fig. 3. Flowchart of the GA optimization algorithm.

by using genetic algorithm (GA) [14,15] as described in the next subsection. Each fuzzy control rule in the controller rule base is of the form If e is E and e is E, then u is U

where E, E and U are the fuzzy sets corresponding to error, change in error and the incremental change in the control output, respectively. In this work, for both the inputs (e and e) and the output ( u), seven fuzzy subsets have been used. These are: PB (positive big), PM (positive medium), PS (positive small), ZE (zero), NS (negative small), NM (negative medium) and NB (negative big).

For each of these fuzzy sets, triangular membership function (MF) has been used. These membership functions have been dened on the common normalized domain [1, 1] and are shown in Fig. 2. From this gure it is observed that the triangles are symmetric with equal base having 50% overlap with neighboring MFs. As each of the two inputs has seven fuzzy sets, there are altogether 49 control rules in the FPIC. The rule base for computing the output u is shown in Table 1 which is a widely used rule base designed with a twodimensional phase plane [16,17]. The control rules in Table 1 are built based on the characteristics of the step response. For example, if the output is falling far away from the set point, a large control signal that pulls the output toward the set point is expected, whereas a small control signal is required when the output is near and approaching the set point. 2.1. Tuning of scaling factors using GA In this work, the scaling factors have been tuned such that the power system oscillations are minimized after a disturbance

Fig. 4. Block diagram of the STFPIC.

S. Hameed et al. / Electric Power Systems Research 78 (2008) 17261735

1729

Fig. 5. MFs for gain updating factor (). ZE: zero; V: very; B: big; M: medium; S: small.

takes place. Specically, the aim is to minimize the error between the actual steady state power owing through the line (Pref ) (in which the TCSC is installed) and the actual power owing through that particular line (Pactual ) following a disturbance. Various performance indices can be used to represent the above goal mathematically. In this paper, the integral of the squared error (ISE) [18] as shown in Eq. (3) has been selected as it tends to place a greater penalty on large errors. This goal can be formulated as the minimization of the objective function F, where
tsim

Table 4 Rule base for e/e NB NM NS ZE PS PM PB NB VB VB VB S VS VS ZE NM VB VB MB SB S S S NS VB B B MB VS MB SB ZE B B VB ZE VB B B PS SB MB VS MB B B VB PM S S S SB MB VB VB PB ZE VS VS S VB VB VB

F=
0

2 ep (t) dt

(3) evaluation is repeated again and again. The algorithm stops when a pre-dened maximum number of generations is achieved. The concepts of reproduction, crossover and mutation are nowadays well known in the literature [14,15] and hence are not described further in this paper. 2.2. Self-tuning fuzzy PI controller (STFPIC) After the scaling factors are found by GA, for enhancing the performance of the FPIC, the output SF is further modulated on-line by a factor , thereby making it a self-tuning FPIC. Essentially, a STFPIC is an adaptive FLC. A FLC is called adaptive if any one of its tunable parameters (scaling factors, membership functions and rules) changes on-line when the controller is being used, otherwise it is a non-adaptive or conventional FLC [16]. An adaptive FLC that ne tunes an already working controller by modifying either its membership functions or scaling factors or both of them is called a self-tuning FLC. The block diagram of the proposed STFPIC is shown in Fig. 4 [13]. Fig. 4 shows that the output SF (gain) of the controller is modied by a self-tuning mechanism (indicated by the dotted boundary). Thus, the output-scaling factor of the self-tuning FLC does not remain xed while the controller is in operation, rather it is modied in each sampling time by the gain updating factor ,

In Eq. (3), ep (t) = Pref Pactual is the error in power ow in the line following a disturbance and tsim is the total time period of simulation. As the objective function of Eq. (3) is non-convex in nature, GA has been used to minimize F. The overall owchart for optimization using GA is shown in Fig. 3. Initially, a number of populations (N) have been generated for the scaling factors. Each of the populations consists of the binary strings corresponding to the scaling factors Ge , G e and Gu . These strings are created in a random fashion with the constraint that the values of Ge , G e and Gu should lie within their specied ranges. The ranges chosen for each of these scaling factors are shown in Table 2 while the parameters used in GA are shown in Table 3. For each of these N sets of values of Ge , G e and Gu , time domain non-linear simulation studies have been carried out for evaluating the objective function F of Eq. (3). For this purpose, a value of 80 s has been chosen for tsim . Based on the values of the objective function, out of these N possible solutions, the good solutions are retained and the others are eliminated (following the principle of survival of the ttest). The selected solutions undergo the processes of reproduction, crossover, and mutation to create a new generation of possible solutions (which are expected to perform better than the previous generation). This process of production of a new generation and its

Fig. 6. TCSC in a two-area four-generator system.

1730

S. Hameed et al. / Electric Power Systems Research 78 (2008) 17261735

Fig. 7. Generators rotor angles with FPIC for base case.

depending on the trend of the controlled process output. The gain updating factor () is determined using fuzzy rules of the form If e is E and e is E then is

From Fig. 4 it is observed that the value of is computed from the normalized values of e and e by a fuzzy rule base. The membership functions used for e and e are exactly same as those used in FPIC. Moreover, the same fuzzy operators as in Fig. 1 have also been used in this case. The membership functions for the factor are dened in the domain [0,1] and are shown in Fig. 5. As each of the two inputs (e and e) to the fuzzy rule base (corresponding to ) has seven fuzzied variables, the rule base has 49 rules for computing the value of . Table 4 shows the rule base for computing . This rule base has been designed to improve the control performance under large disturbances such as three-phase short circuit on the transmission lines, a sudden loss of generating unit or a large loss of load, etc. For example, immediately after a large disturbance, e may be small but e will be sufciently large (they will be of same sign) and, for this case, is supposed to be large to increase the gain. Therefore, under these circumstances, the appropriate rules are IF e is PS and e is PM THEN is B or IF e is NS and e is NM THEN is B. On the other hand, for steady state conditions (i.e., e 0 and e 0), controller gain should be very small (e.g., IF e is ZE and e is ZE THEN is ZE) to avoid chattering problem around the set point. Further justication for using the rule base in Table 4 can be found in [12]. The principal steps for STFPIC can be summarized as follows: Step 1: Tune the SFs of the STFPIC without the gain tuning mechanism and assuming = 1 (i.e., conventional FLC) for a given process to achieve a reasonably good control performance. Here, genetic algorithm [19] has been used for tuning the conventional FLC. At the end of this step, we get a good controller without selftuning and this controller becomes the starting point (input) for the self-tuning controller in Step 2.

Step 2: Following [12], set the output SF (Gu ) of the self-tuning FLC K times greater than that obtained in Step 1 keeping the values of Ge and G e same as those of the conventional FLC. In this step = 1, which is obtained from the rule base in Table 4. This enhancement of Gu for the STFPI is found empirically [12] with an objective to improve the control performance.

3. Case studies In this work, the effectiveness of the proposed self-tuning FPIC has been validated on two different multi-machine power systems: (a) 2 area 4 machine system and (b) 10 machine 39 bus system. In the next subsection, the details of these two systems are presented.

Fig. 8. Active power ow in line 1110 and TCSC capacitive reactance with FPIC for base case.

S. Hameed et al. / Electric Power Systems Research 78 (2008) 17261735

1731

Fig. 9. Rotor angle of generator 2 with STFPIC for base case.

Fig. 10. Active power ow in line 1110 and TCSC capacitive reactance with STFPIC for base case.

3.1. Study systems The schematic diagram of the rst study system of this work, i.e. the two area four machine system is shown in Fig. 6. The detailed data of this system can be found in [20]. In this system, machines 1 and 2 form a coherent group, and machines 3 and 4 form the other coherent group. There are three tie lines connecting the two coherent areas. As shown in Fig. 6, a TCSC has been assumed to be installed in one of these tie lines. TCSC is inserted in the middle of one tie line. Due to lack of space, the one line diagram of the 10 machine system is not shown in this paper. However, the data of this system has been taken from [21] and is given in the Appendix A for ready reference. In this system, a TCSC has been assumed

to be installed in the middle of the line connecting buses #39 and #36. 3.2. System modelling In this work, the synchronous generator has been represented by a eld circuit on the d-axis and one equivalent damper winding on the q-axis. The machine differential equations and the differential equation for the static exciter for the ith machine (sufx i is not used in these equations just for simplicity) are given below. The system loads are represented by constant impedances. d = s dt (4)

Fig. 11. Generators rotor angles with FPIC for 10% increased loading.

1732

S. Hameed et al. / Electric Power Systems Research 78 (2008) 17261735

Fig. 12. Active power ow in line 1110 and TCSC capacitive reactance with FPIC for 10% increased loading.

Fig. 14. Active power ow in line 1110 and TCSC capacitive reactance with STFPIC for 10% increased loading.

1 d = [D( s ) + Tm Te ] 2H dt dEq dt dEd dt dEfd dt = = 1 [Eq + (xd xd )id + Efd ] Td0 1 [Ed (xq xq )iq ] Tq0 1 [Efd + KA (Vref Vt )] TA

(5) (6)

KF Efd RF dRF + = TF dt (TF )2 The electrical torque, Te is expressed as follows Te = Ed id + Eq iq + (xd xq )id iq

(11)

(12)

(7)

The notations used in the above Eqs. (412) are quite standard and hence they are not dened in this paper. For more details, the readers are suggested to refer [21]. 3.3. Simulation results In this paper, the effectiveness of the proposed FPIC and STFPIC controllers has been studied through detailed non-linear time domain simulation studies under three phase, ve cycle, solid short circuit faults. The short circuit faults have been assumed to occur at t = 5 s. The simulation studies have been carried out in the MATLAB/SIMULINK environment [22]. For illustrating the efcacy of the fuzzy controllers developed in this work, results pertaining to three different situations are presented. These situations are: (a) study system without any TCSC, (b) study system with a TCSC controlled by FPIC and (c) study system with a TCSC controlled by STFPIC. The simulation results pertaining to these three cases are

(8)

However, IEEE Type I exciter has been used for 10-machine 39bus system. The differential equations for this exciter are dEfd dt = KE + SE (Efd ) TE Efd + VR TE (9)

where SE (Efd ) = Aex eBex Efd VR KA RF KA KF KA (Vref Vt ) dVR = + E + TA TA TA TF fd TA dt (10)

Fig. 13. Rotor angle of generator 2 with STFPIC for 10% increased loading.

Fig. 15. Rotor angle of generator 3 for 75% increased loading.

S. Hameed et al. / Electric Power Systems Research 78 (2008) 17261735

1733

Fig. 16. Active power ow in line 3936 and TCSC capacitive reactance with FPIC for 75% increased loading.

Fig. 18. Active power ow in line 3936 and TCSC capacitive reactance with STFPIC for 75% increased loading.

Fig. 17. Rotor angle of generator 3 for 75% increased loading with STFPIC.

presented below for both the study systems. In the following gures, the machine rotor angles have been measured with respect to the centre of inertia (COI) [21]. 3.3.1. Two area four machine system In this system, the short circuit fault has been assumed to take place at bus 9. The variations of the machine rotor angles and active power ow in line 11 10 (P1110 ) for the base case loading condiTable A1 Machine data for 39 bus system Parameter M/C 1 M/C 2 M/C 3 M/C 4 M/C 5 M/C 6 M/C 7 M/C 8 M/C 9 M/C 10 xd (p.u.) 0.295 0.02 0.2495 0.33 0.262 0.254 0.295 0.29 0.2106 0.2 xd (p.u.) 0.0647 0.006 0.0531 0.066 0.0436 0.05 0.049 0.057 0.057 0.004 Tdo (s) 6.56 6 5.7 5.4 5.69 7.3 5.66 6.7 4.79 5.7 xq (p.u.) 0.282 0.019 0.237 0.31 0.258 0.241 0.292 0.28 0.205 0.196

tion (the loading condition as described in [20]) are displayed in Figs. 7 and 8(a) corresponding to scenario (a) and (b), respectively. In Fig. 7, the dotted lines show the variations of rotor angles without (w/o) TCSC whereas the variations of angles with TCSC FPIC are depicted with solid lines. From Figs. 7 and 8(a), it is observed that there are substantial oscillations in the system (without any TCSC), which is damped to a large extent by the proposed TCSC FPIC. The variation of the TCSC reactance (XTCSC ) is shown in Fig. 8(b) for ` situation (b). The performance of the proposed TCSC STFPIC vis-avis that obtained with TCSC FPIC is depicted in Figs. 9 and 10. For implementing STFPIC, a value of 3 has been chosen for K. In Fig. 9, the variation of the rotor angle of machine 2 are shown whereas in Fig. 10(a) and (b), the variations of P1110 and XTCSC are shown respectively. From these gures it is observed that application of TCSC STFPIC improves the system damping further as compared to TCSC FPIC. To investigate the performance of the proposed TCSC controller at enhanced loading condition, simulation studies were carried out by increasing the system loads by 10% from the base case loading condition. The results are shown in Figs. 11 and 12 corresponding to scenario (a) and (b), respectively. From Figs. 11 and 12(a) it is observed that at 10% higher loading, the system is unstable without TCSC, which is made stable with acceptable level of damping by the proposed TCSC FPIC. Fig. 12(b) shows the variation of the TCSC reactance for this case with TCSC FPIC only. The performance of the TCSC STFPIC for this loading condition is shown in Figs. 13 and 14. Again, from these gures it is observed that the system damping improves further (as compared to that obtained by TCSC FPIC), by using TCSC STFPIC.

xq (p.u.) 0.0647 0.006 0.0531 0.066 0.0436 0.05 0.049 0.057 0.057 0.004

Tqo (s) 1.5 0.7 1.5 0.44 1.5 0.4 1.5 0.41 1.96 0.5

H (s) 30.3 500 35.8 26 28.6 34.8 26.4 24.4 34.5 42

D (p.u.) 0 0 0 0 0 0 0 0 0 0

x1 (p.u.) 0.0518 0.0048 0.0425 0.0528 0.0349 0.04 0.0392 0.0456 0.0456 0.0032

1734 Table A2 Exciter data for 39 bus system KA (p.u.) Ex1 Ex2 Ex3 Ex4 Ex5 Ex6 Ex7 Ex8 Ex9 Ex10 6 20 5 40 5 5 40 5 40 25 TA (s) 0.05 0.2 0.06 0.02 0.06 0.02 0.02 0.02 0.02 0.06

S. Hameed et al. / Electric Power Systems Research 78 (2008) 17261735

KE (p.u.) 0.63 1.0 0.02 1.0 0.05 0.04 1.0 0.05 1.0 0.02

TE (s) 0.41 0.314 0.5 0.73 0.5 0.47 0.73 0.53 1.4 0.50

Efd min (p.u.) 6 6 6 6 6 6 6 6 6 6

Efd max (p.u.) 6 6 6 6 6 6 6 6 6 6

KF (p.u.) 0.25 0.063 0.08 0.03 0.08 0.075 0.03 0.085 0.03 0.08

TF (s) 0.5 0.35 1.0 1.0 1.0 1.25 1.0 1.26 1.0 1.0

Aex (p.u.) 0.705 0 0.0184 0 0.0035 0.0021 0.493 0.0028 0.61 0

Bex (p.u.) 0.288 0 0.625 0 0.82 0.857 0.311 0.837 0.3 0

3.3.2. Ten machine 39 bus system In this system, the short circuit fault has been assumed to take place at bus 24. A large number of simulation studies have been carried out at various increased system loading conditions to investigate the suitability of the proposed TCSC controller for enhancing the power carrying capacity of the system. From these studies it has been found that the proposed TCSC controller helps to increase

Table A3 Line data for 39 bus system Line no. Bus From 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 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 22 16 20 39 39 32 32 30 29 25 23 12 37 37 36 36 36 36 35 34 33 28 26 26 26 25 23 22 21 20 20 19 18 17 16 16 15 15 14 14 13 13 12 12 11 11 To 6 1 3 30 5 33 31 4 9 8 7 10 27 38 24 21 39 37 36 35 34 29 29 28 27 26 24 23 22 33 31 2 19 18 31 17 18 16 34 15 38 14 25 13 12 2 Impedance R (p.u.) 0 0 0 0.0007 0.0007 0.0016 0.0016 0.0009 0.0008 0.0006 0.0005 0 0.0013 0.0007 0.0003 0.0008 0.0016 0.0007 0.0009 0.0018 0.0009 0.0014 0.0057 0.0043 0.0014 0.0032 0.0022 0.0006 0.0008 0.0004 0.0004 0.0010 0.0023 0.0004 0.0007 0.0006 0.0008 0.0002 0.0008 0.0008 0.0011 0.0013 0.0070 0.0013 0.0035 0.0010 X (p.u.) 0.0143 0.0250 0.0200 0.0138 0.0142 0.0435 0.0435 0.0180 0.0156 0.0232 0.0272 0.0181 0.0173 0.0082 0.0059 0.0135 0.0195 0.0089 0.0094 0.0217 0.0101 0.0151 0.0625 0.0474 0.0147 0.0323 0.0350 0.0096 0.0135 0.0043 0.0043 0.0250 0.0363 0.0046 0.0082 0.0092 0.0112 0.0026 0.0129 0.0128 0.0133 0.0213 0.0086 0.0151 0.0411 0.0250 0 0 0 0 0 0 0 0 0 0 0 0 0.1608 0.06595 0.0340 0.1274 0.1520 0.0671 0.0855 0.1830 0.08615 0.1245 0.5145 0.3901 0.1198 0.2565 0.1805 0.0923 0.1274 0.03645 0.03645 0.6000 0.1902 0.0390 0.06945 0.0565 0.0738 0.0217 0.0691 0.0671 0.1069 0.1107 0.0730 0.1286 0.34935 0.3750 B/2 (p.u.)

the system power carrying capacity quite signicantly. As it is not possible to include all the simulation results in the paper due to lack of space, a few representative results are presented below to illustrate the effectiveness of the proposed TCSC fuzzy controller for enhancing the power carrying capacity of the system under study. The performance of the proposed TCSC FPIC at 75% enhanced loading condition (from the base case loading condition as described in [21]) is shown in Figs. 15 and 16. From these gures it is observed that at 75% increased loading condition, the system becomes unstable upon occurrence of the fault (without any TCSC) while the proposed TCSC FPIC is able to stabilize the system. The operation of the TCSC STFPIC at 75% enhanced loading condition is shown in Figs. 17 and 18. For implementing STFPIC, a value of 10 has been chosen for K. From these gures (Figs. 17 and 18) it

Table A4 Bus data for 39 bus system Bus 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 27 28 29 30 31 32 33 34 35 36 37 38 39 Type Swing PV PV PV PV PV PV PV PV PV PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PQ PL (p.u.) 0.0920 11.0400 0 0 0 0 0 0 0 0 0 0 3.2200 5.0000 0 0 2.3380 5.2200 0 0 2.7400 0 2.7450 3.0860 2.2400 1.3900 2.8100 2.0600 2.8350 6.2800 0 0.0750 0 0 3.2000 3.2940 0 1.5800 0 QL (p.u.) 0.0460 2.5000 0 0 0 0 0 0 0 0 0 0 0.0240 1.8400 0 0 0.8400 1.7600 0 0 1.1500 0 0.8466 0.9220 0.4720 0.1700 0.7550 0.2760 0.2690 1.0300 0 0.8800 0 0 1.5300 0.3230 0 0.3000 0 PG (p.u.) 5.4282 10.000 6.5000 5.0800 6.3200 6.5000 5.6000 5.4000 8.3000 2.5000 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 QG (p.u.) 1.5724 2.2624 1.6606 1.5510 0.8381 2.8105 2.2967 0.2757 0.5970 1.8388 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

S. Hameed et al. / Electric Power Systems Research 78 (2008) 17261735

1735

is clearly observed that the proposed TCSC STFPIC improves the system stability further as compared to TCSC FPIC. 4. Conclusion In this paper, a self-tuning fuzzy logic controller has been proposed for TCSC to improve the power system damping. The effectiveness of the proposed STFPIC has been validated on two multi-machine power systems through detailed non-linear simulation studies under wide variation of operating conditions. Also the performance of STFPIC has been compared to that obtained with the standard FPIC. From the simulation studies it has been observed that both FPIC and STFPIC help to enhance the system power carrying capability quite signicantly. However, the performance of the proposed STFPIC for damping oscillations is better than that obtained by the standard FPIC and the superiority of the proposed STFPIC over the standard FPIC becomes more pronounced at higher loading conditions of the power system. Appendix A The data for 39 bus system are given in Tables A1A4 below. List of symbols e(k) error at kth sample e(k) change in error eN normalized error normalized change in error eN ep (t) error in power ow in the line following a disturbance scaling factor corresponding to error Ge G e scaling factor corresponding to change in error Gu scaling factor corresponding to control output Pactual actual power owing through the line following a disturbance Pref reference or steady state power owing through the line tsim total time period of simulation uN normalized incremental change in control output u(k) control output at kth sample u(k) actual incremental change in control output at kth sample ysp set-point or desired system output y(k) actual system output at kth sample

References
[1] N.G. Hingorani, L. Gyugyi, Understanding FACTS: Concepts and Technology of Flexible AC Transmission System, IEEE Press, 2000. [2] E.V. Larsen, J.J. Sanchez-Gasca, J.H. Chow, Concepts for design of FACTS controllers to damp power swings, IEEE Trans. Power Syst. 10 (2) (1995) 948956. [3] R.M. Mathur, R.K. Verma, Thyristor-based FACTS Controllers for Electrical Transmission Systems, IEEE press, Piscataway, 2002. [4] X. Zhou, J. Liang, Overview of control schemes for TCSC to enhance the stability of power systems, IEE Proc. C, Generat. Transm. Distrib. 146 (2) (March, 1999) 125130. [5] D.Z. Fang, Y. Xiaodong, S. Wennan, H.F. Wang, Oscillation transient energy function applied to the design of a TCSC fuzzy logic damping controller to suppress power system inter-area mode oscillations, IEE Proc. C. Generat. Transm. Distrib. 150 (2) (March, 2003) 233238. [6] Xiaobo. Tan, Naiyao. Zhang, Luyuan. Tong, Zhonghong. Wang, Fuzzy control of thyristor-controlled series compensator in power system transients, Fuzzy Sets Syst. 110 (3) (2000) 429436. [7] P.K. Dash, S. Mishra, G. Panda, Damping multimodal power system oscillation using hybrid fuzzy controller for series connected FACTS devices, IEEE Trans. Power Syst. 15 (4) (2000) 13601366. [8] P.K. Dash, S. Mishra, Damping of multimodal power system oscillations by FACTS devices using non-linear Takagi-Sugeno fuzzy controller, Electric. Power Energy Syst. 25 (6) (2003) 481490. [9] Chun-Feng. Lu, Chia-Feng. Juang, Evolutionary fuzzy control of exible AC transmission system, IEE Proc. C, Generat. Transm. Distrib. 152 (4) (2005) 441 448. [10] Laiq. Khan, K.L. Lo, Hybrid micro-GA based FLCs for TCSC and UPFC in a multimachine environment, Electric Power Syst. Res. 76 (9/10) (2006) 832843. [11] A. Kazemi, M.V. Sohrforouzani, Power system damping using fuzzy controlled facts devices, Electric. Power Energy Syst. 28 (5) (2006) 349357. [12] R.K. Mudi, N.R. Pal, A robust self-tuning scheme for PI- and PD-type fuzzy controllers, IEEE Trans. Fuzzy Syst. 7 (1) (1999) 216. [13] K. Pal, R.K. Mudi, N.R. Pal, A new scheme for fuzzy rule-based system identication and its application to self-tuning fuzzy controllers, IEEE Trans. Syst. Man, Cybernet. B: Cybernet. 32 (4) (2002) 470482. [14] D.E. Goldberg, Genetic algorithms in search, in: Optimization and Machine Learning, Pearson Education Asia, 2001. [15] T.J. Ross, Fuzzy Logic with Engineering Applications, 2nd ed., John Wiley & Sons, Ltd., 2005. [16] D. Driankov, H. Hellendoorn, M. Reinfrank, An Introduction to Fuzzy Control, Springer-Verlag, New York, 1993. [17] Han-Xiong Li, H.B. Gatland, Conventional fuzzy control and its enhancement, IEEE Trans. Syst. Man, Cybernet. B: Cybernet. 26 (5) (1996) 791796. [18] Y. Zhao, E.G. Collins Jr., Fuzzy PI control design for an industrial weigh belt feeder, IEEE Trans. Fuzzy Syst. 11 (3) (2003) 311319. [19] A. Popov, Genetic Algorithm for Optimization. Web-site: http://automatics. hit.bg/. [20] P. Kundur, Power System Stability and Control, McGraw-Hill Inc., New York, 1994. [21] K.R. Padiyar, Power System DynamicsStability and Control, Interline Publishing Ltd., Bangalore, India, 1996. [22] MATLAB/SIMULINK Software, Version 7.0.1, Release 14, The Math Works Inc.