Optimal Location and Parameters Setting of UPFC based on GA and PSO for Enhancing Power System Security under

Single Contingencies
H. I. Shaheen, Student Member, IEEE, G. I. Rashed, Student Member, IEEE, S. J. Cheng, Senior Member, IEEE
meet the N-1 security criterion which is somewhat conservative and costly. An alternative solution to improve the security of power system is the Flexible AC Transmission Systems (FACTS) devices which is a concept proposed by N.G.Hingorani [1]. FACTS devices can improve the stability of the power network, reduce the flows of heavily loaded lines by controlling their parameters, and maintain the bus voltages at desired levels. Consequently, they can improve the power system security under contingencies. Unified Power Flow Controller (UPFC) is a versatile FACTSs device which can independently or simultaneously control the active power, the reactive power, and the bus voltage to which it is connected [2]. This controller offers substantial advantages for the static and dynamic operation of power system. However, to achieve such functionality, it is important to determine the optimal location of this device to be installed in power system with appropriate parameters setting. Since UPFC can be installed in different locations, its effectiveness will be different. Therefore, we will face the problem of where we should install the UPFC. For this reason, some performance indices must be satisfied. The following factors can be considered in the selection of the optimal installation and the optimal parameters setting of UPFC: The topology of the system, the stability margin improvement, the power transmission capacity increasing, and the power blackout prevention. Therefore, conventional power flow algorithm [3] should incorporate with UPFC considering one or all of the above mentioned factors. In the last decade, new algorithms have been developed for the optimal power flow incorporating with UPFC device as well as for the optimal placement of UPFC. Some of them are: A sensitivity-based approach which has been developed for finding suitable placement of UPFC [4], an evolutionary programming approach to determine the optimal allocation of multi-type FACTS devices [5] and [6], a Genetic Algorithm which proposed for solving the optimal location of FACTS [7] and [8], and a particle swarm technique for optimal location of FACTS devices [9] and [10]. Also a lot of work has been done in the contingency analysis area. Different contingency selection methods can be found in [11] - [14]. Operation scheme of FACTS devices to enhance the power system steady-state security level considering a line

Abstract Unified Power Flow Controller (UPFC) is one of the most effective Flexible AC Transmission Systems (FACTS) devices for enhancing power system security. However, to what extent the performance of UPFC can be brought out, it highly depends upon the location and the parameters setting of this device in the system. This paper presents an approach to find out the optimal placement and the optimal parameters setting of UPFC for enhancing power system security under single contingencies (N-1 Contingency). Firstly, we perform a contingency analysis and ranking process to determine the severest line tripping contingencies considering line overloads and bus voltage violations as a performance index. Secondly, we apply Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) techniques to find out the optimal location and the optimal parameters setting of UPFC corresponding to the determined contingencies scenarios. To verify our proposed approach, we perform simulations on an IEEE 6-bus and an IEEE 14-bus power systems. The results weve obtained indicate that GA and PSO can successfully achieve the proper settings and placement of UPFC. Installing UPFC in such location can significantly enhance the security of power system by eliminating or minimizing the overloaded lines and the bus voltage violations. Index Terms Contingency Analysis, Power Flow, Unified Power Flow Controller (UPFC), Genetic Algorithm (GA), Particle Swarm Optimization (PSO).

I. INTRODUCTION

HE secure operation of power system has become an important and critical issue in todays large, complex, and load demand-increasing power systems. Security constraints such as thermal limits of transmission lines and bus voltage limits must be satisfied under any certain operating point. Commonly, Power systems are planned and operated based on the N-1 security criterion, which implies that the system should remain secure under all important first contingencies. One solution to cope with this problem is to design the system to

H. I. Shaheen is a PhD candidate in College of Electrical and Electronic Engineering, Huazhong University of Science and Technology (HUST), china. (e-mail: shaheen@ieee.org ). G. I. Rashed is a PhD. candidate in College of Electrical and Electronic Engineering, Huazhong University of Science and Technology (HUST), china. (e-mail: gh197493@yahoo.com ). S. J. Cheng is a life professor in college of Electrical and Electronic engineering, Huazhong University of Science and Technology (HUST), China. (e-mail:Sjcheng@hust.edu.cn).

2008 IEEE.

contingency analysis is suggested in [15]. A method for contingency selection and security enhancement of power systems by optimal placement of FACTS devices using GA is presented in [16]. This paper deals with the application of Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) to find out the optimal location and the optimal parameters setting of UPFC device under single contingencies to eliminate or minimize the overloaded lines and bus voltage violations. The rest of this paper is organized as follows: In section II, we describe the contingency analysis and ranking procedure as well as the objective function which we adopt in this work. In section III, we describe our implementation of GA and PSO algorithms according to the optimization objective. Section IV presents the simulation results from implementing GA and PSO algorithms in the MATLAB environment. In this section, we discuss two case studies for an IEEE 6-bus system and an IEEE 14-bus system. Finally, in section V, we conclude the main results and achievements that we have obtained in this paper. II. PROBLEM FORMULATION A. Contingency Analysis Procedure A contingency is considered to be the outage of a generator, a transformer or a line. The system may become unstable and enters an insecure state when a contingency event is occurred. Contingency analysis is one of the most important functions performed in power systems to establish appropriate preventive and / or corrective actions for each contingency. Contingency analysis procedure consists of line contingency analysis, contingency selection, detection of overloaded lines and bus voltage violations, and ranking of the severest contingencies cases. In this paper we focus only on the single contingencies result from line outage (N-1 Contingency). For each line outage contingency in the system, we list the all overloaded lines and the buses which have voltage violations, and then the lines are ranked according to the severity of the contingency, in other words, according to the number of the thermal and voltage violations limits. Then the most critical contingencies are determined. After determining the most critical contingencies scenarios, GA and PSO techniques are applied to find the optimal location and parameters setting of UPFC. Installing UPFC in such optimal location with such optimal parameters will eliminate or minimize the overloaded lines and the bus voltage violations under these critical contingencies according to the objective function described in B. B. Objective Function As the main objective of this work is to determine the optimal location and the optimal parameters setting of the UPFC in the power network to eliminate or minimize the overloaded lines and the bus voltage violations under the most critical single contingencies, the following performance index is selected:

Min Ft Subject to: g ( x, u ) = 0

h ( x, u ) 0

(1) (2) (3)

Where: Ft represents the objective function to be minimized; g ( x, u ) represents the equality constraints corresponding to active and reactive power balance equations; h ( x, u ) represents the inequality constraints corresponding to UPFC parameter bounds limits, active and reactive power generation limits, bus voltage limits, and phase angles limits; x represents the state vector of the power system consisting of voltage magnitude and phase angles; and u represents the vector of control variables to be optimized (i.e., location of UPFC and its parameters setting: VcR and
VvR). For enhancing the system security level, UPFC should be located in order to eliminate or minimize the line overloads and to prevent the bus voltages violations. Therefore, we considered the following technical objective function [17]: 2r ntl S 2 q nb Vmref V l m Ft = wl w + (4) m V mref l =1 Sl max m=1

Where: Sl and Sl max represent the current apparent power in line l and the apparent power rate of line l , respectively;
Vm represents the voltage magnitude at bus m ;

Vmref represents the bus m nominal voltage; wl and wm represent two weights and are determined in order to have the same index value for 10% voltage difference and for 100% branch loading; q and r represent two coefficients are used to penalize more or less overloads and voltage variations, respectively (for the presented study, they are considered to be equal to 2); and ntl and nb represent the number of lines and the number of buses in the system.

III. METHODOLOGIES FOR OPTIMAL LOCATION OF UPFC

A. Description of the Implemented Genetic Algorithm technique

Genetic algorithm is a kind of stochastic search techniques based on the mechanism of natural selection and survival of the fittest [18] and [19]. Further, it combines the function evaluation with the randomized and/or well-structured exchange of information among the solutions to arrive at a global optimum. More importantly, GA appears attractive because of its superior robust behavior in nonlinear environments over the other optimization techniques. The architecture of the GA implementation can be segregated into the following three constituent phases namely: initial

population generation, operations. 1)

fitness

evaluation

and

genetic

A Brief Outline of GA Computational Tasks The GA control parameters, such as population size, crossover probability and mutation probability are selected, and an initial population of floating strings of finite length is randomly generated. Each of these individuals, comprising a number of chromosomes, represents a feasible solution to the search problem. Basically, average minimum and maximum fitness of all individuals within a generation are computed. If a pre-defined convergence criterion is not satisfied, then the genetic operations comprising selection, crossover and mutation are carried out. Fundamentally, the selection mechanism attempts to apply pressure upon the population in a manner similar to that of the natural selection found in biological systems. A new population is created with worse performing individuals eliminated whilst the most highly fit members in a population are selected to pass on information to the next generation. In this work, the selection function called deterministic sampling selection is adopted. The method ensures that the bigger fitness individuals are remaindered into the next generation. Conceptually, pairs of individuals are randomly chosen from the population and the fit of each pair is allowed to mate. Each pair of mates creates a child having some mix of the two parents. 2) An Advanced Computational Refinement of GA: The crossover previously mentioned is the kernel of genetic operations. It promotes the exploration of new regions in the search space using randomized mechanism of exchanging information between strings. Two individuals previously placed in the mating pool during reproduction are randomly selected. A crossover point is then randomly selected and information from one parent up to the crossover point is exchanged with the other parent. This is specifically illustrated in Fig. 1 for the implemented simple crossover technique, which is adopted in this work.

exploration of the search space. This unique feature of PSO overcomes the premature convergence problem and enhances the search capability. Also unlike the other methods, the solution quality doesnt rely on the initial population. Starting anywhere in the search space, the PSO algorithm ensures the convergence to the optimal solution. In the following paragraph we briefly introduce PSO mechanism. More detail can be found in [23]. PSO is similar to the evolutionary computation techniques in that, a population of potential solutions to the optimal problem under consideration is used to probe the search space. Each potential solution is also assigned a randomized velocity and the potential solutions, called particles, correspond to individuals. Each particle in PSO flies in the D-dimensional problem space with a velocity dynamically adjusted according to the flying experiences of its own and other particles. The ith particle is represented location of the as X i = [ xi1, xi 2 , , xiD ] , where, xid [ld , ud ] , d [1, D ] .

and ud , are the lower and upper bounds for dth dimension, respectively. The best previous position (which gives the best fitness value) of the ith particle is recorded and represented
l d

as Pi = [ Pi1, Pi 2 , , PiD ] , which is also called Pbest . The index of the best particle among the all particles in the population is represented by the symbol g . The location Pg is also denoted as gbest . The velocity of the ith particle is represented as Vi = [vi1, vi 2 , , viD ] and is clamped to a maximum velocity Vmax = [vmax1 , vmax2 , , vmax D ]

which is specified by the user. At each time step, the particle swarm optimization concept regulates the velocity and location of each particle toward its Pbest and gbest locations according to the equation (5) and (6), respectively. n +1 n n n n n n vid = wvid + c1r 1 ( Pid xid )+c2 r2 ( Pgd n+1 n n+1 xid = xid + vid i =1,2, , m , d = 1, 2, , D Where: c1 and c 2 are two positive constants, called
n xid )

(5) (6)

Fig. 1. Simple crossover

Another process also considered in this work is the mutation process of randomly changing encoded bit information for a newly created population individual. Mutation is generally considered as a secondary operator to extend the search space and cause escape from a local optimum when used prudently with the selection and crossover schemes. B. Description of the Implemented Particle Swarm Optimization Technique Particle Swarm Optimization (PSO) is a stochastic global optimization approach and its main strength is in its simplicity and fast convergence rates [20]-[22]. Unlike Genetic Algorithm (GA) and other heuristic algorithms, PSO has the flexibility to control the balance between the global and local

cognitive and

social parameters respectively, m is the size of the swarm, D is the number of members in a particle, r 1 and r2 , are random numbers, uniformly distributed in [0,1] , n is the pointer of iterations (generations), and w is the inertia weight, which provides a balance between global and local explorations and thus requiring less iteration on average to find a sufficiently optimal solution. It is specified by equation (4): wmax wmin iter w = wmax (7) itermax Where: wmax is the initial weight, wmin is the final weight,

iter is the current iteration number, and itermax is the maximum iteration number (generations).

uniform mutation have been adopted for the implemented GA. The other GA parameters are presented in Table I.
TABLE I PARAMETER VALUES FOR GA Parameter of GA No. of variables Length of individual Population size (np) of individuals Maximum number of generations 3 4 30

IV. SIMULATION RESULTS A. Simulation Tool and Power Systems Matlab Codes for GA, PSO, and a modified power flow algorithm to include UPFC were developed and incorporated together for the simulation purposes. To investigate the validation of the applied techniques, both GA and PSO algorithms have been tested on the following two test systems: An IEEE 6-bus system (shown in Fig. 2), and an IEEE 14-bus system (shown in Fig. 3). The data for the above mentioned systems are taken from [24] and [25], respectively. Simulations were done on Pentium IV, 2.66 GHz, and 512MB of RAM.

g max

100 1

Number of offspring per pair of parents

For PSO technique the parameters utilized in this simulation are shown in Table II.
TABLE II PARAMETER VALUES FOR PSO Parameters of PSO

c1 and c 2

1.5 0.9 0.4 10 30 100

wmax wmin
deviation of initial velocity number of swarm beings number of flights

The results obtained from applying GA and PSO techniques for both systems are presented as follows: 1. IEEE 6-Bus Test System This system consists of three generators, six buses, eleven transmission lines, and three loads as shown in Fig. 1. Contingency analysis and ranking process is performed on this system. There are 11 possible single contingencies. For each single line outage, we find the Number of Over Loaded Lines (NOLL) and the Number of Voltage Violation Buses (NVVB). Then we rank the lines according to the severity of the contingency, in other words, according to the Performance Index PI = ( NOLL + NVVB ) as shown in Table III. PI is zero for the remaining lines.
TABLE III RANKING OF BRANCHES FOR IEEE 6-BUS SYSTEM Line Number Number of Over Loaded Lines (NOLL) Line To Bus From Bus Number of Voltage Violation Buses (NVVB) Performance Index PI = ( NOLL + NVVB )

Fig. 2. The IEEE 6-bus system

Fig. 3. The IEEE 14-bus system

B. Observed Results In this paper, the optimal location and the optimal parameters setting of UPFC device in the network were optimized to eliminate or minimize the line overloads and the bus voltage violations under single contingencies. Therefore, the UPFC is modeled for the purpose of fundamental steady-state analysis as a coordination of two synchronous voltage sources, VcR and VvR, respectively [26]. The applied GA and PSO algorithms were implemented to find the optimal values of the above mentioned variables. Simple Crossover, deterministic sampling selection, and non

1 2 6 9 5 4

1 1 2 3 2 2

4 5 4 6 6 5

3 3 2 2 2 1

1 0 1 1 0 0

4 3 3 3 2 1

From Table III, we can find that lines (1, 2,6,9,5, and 4) are the worst contingencies scenarios in this system. In the case of one of this lines is outage, the most overloaded lines and bus voltage violations will be encountered. Therefore, to investigate the effectiveness of UPFC on the system under such

Rank 1 2 3 4 5 6

cases, we apply both GA and PSO techniques to find the optimal location and the optimal parameters setting of UPFC which eliminate or minimize the overloaded lines and bus voltage violations. Overloaded lines and bus voltage violations before and after placing UPFC in this system with optimal location and optimal parameters setting of UPFC obtained by applying both GA and PSO techniques are shown in Table IV and Table V, respectively.
TABLE IV OVERLOADED LINES AND BUS VOLTAGE VIOLATIONS BEFORE AND AFTER PLACING UPFC FOR IEEE 6-BUS SYSTEM WITH OPTIMAL LOCATION AND OPTIMAL PARAMETERS SETTING OF UPFC BY GA Tripped After Placing Before Placing UPFC Line UPFC Optimal Placement of UPFC Overloaded Lines Voltage Violation Buses Voltage Violation Buses Overloaded Lines Overloading % Overloading % Optimal Setting of UPFC Rank

From Tables IV and V, we can find that all bus voltage violations are eliminated by placing UPFC in an optimal location with optimal parameters setting by both GA and PSO. While for overloaded lines, despite that placing UPFC in the system didnt result in eliminating all of them, but most of overloaded lines are eliminated and the power flow distribution in the rest overloaded lines is significantly reduced. Fig. 4 shows the power flow distribution in the IEEE 6-bus system when line 2 is outage. Before placing UPFC in this system, there were 3 overloaded lines, and after placing UPFC two of them are eliminated and the power flow distribution for the third one is reduced. These results are achieved by applying both GA and PSO techniques.

From Bus

To Bus

VcR

VvR
1.1 1

1 2 3 2 2

5 4 6 6 5

1-5 1-2 2-4 1-4 1-2 3-5 1-4 3-5 2-6 3-5 2-3 3-6 3-5

106.64 122.24 Bus_4 116.97 2-4 113.1 111.1 115.15 1-4 108.06 103.96 158.95 1-4 116.8 Bus_4 104.74 129.48 Bus_6 144.28 138.14 114.42 102.43 -

11 0.05068

10 0.103252 1.0999 2 10 0.088986 1.0596 3 11 0.106943 1.0999 4 11 0.066484 1.0999 5 2 0.111071 1.0999 6 Fig. 4. Power flow distribution for IEEE 6-bus system when line 2 is outage by GA and PSO

Fig.5 shows the minimization of the objective function achieved by both GA and PSO techniques when line 2 is outage. Because of the limited space for this paper, we only choose this case as a sample of the results weve obtained.
9.30 9.20 9.10 O bjective Function 9.00 8.90 8.80 8.70 8.60 GA PSO

TABLE V OVERLOADED LINES AND BUS VOLTAGE VIOLATIONS BEFORE AND AFTER PLACING UPFC FOR IEEE 6-BUS SYSTEM WITH OPTIMAL LOCATION AND OPTIMAL PARAMETERS SETTING OF UPFC BY PSO Tripped After Placing Before Placing UPFC Line UPFC Optimal Placement of UPFC Overloaded Lines Voltage Violation Buses Voltage Violation Buses Overloaded Lines Overloading % Overloading % Optimal Setting of UPFC Rank

From Bus

To Bus

VcR

VvR

8.50 8.40

1 2 3 2 2

5 4 6 6 5

1-5 1-2 2-4 1-4 1-2 3-5 1-4 3-5 2-6 3-5 2-3 3-6 3-5

106.64 122.24 Bus_4 116.97 2-4 114.3 113.4 115.15 1-4 108.06 103.96 158.95 1-4 123.8 Bus_4 104.74 129.48 Bus_6 144.28 138.14 114.42 102.43 -

10

20

30

40

2 0.108185 1.0304 1

50 60 Generation

70

80

90

100

Fig. 5. Minimization of the objective function by both GA and PSO techniques for IEEE 6-bus system when line 2 is outage.

10 0.103587

1.1

10 0.131492 1.0291 3 11 0.108366 1.0564 4 11 0.090063 1.0361 5 10 0.077639 1.0913 6

2. IEEE 14-Bus Test System This system consists of five generators, fourteen buses, twenty transmission lines, and eleven loads as shown in Fig. 2. Contingency analysis and ranking process is also performed on this system as mentioned in IEEE 6-bus system. There are 20 possible single contingencies. For each single line outage, we find the Number of Over Loaded Lines (NOLL) and the Number of Voltage Violation Buses (NVVB). Then we

rank the lines according to the severity of the contingency, in other words, according to the Performance Index PI = ( NOLL + NVVB ) as shown in Table VI. PI is zero for the remaining lines.
TABLE VI RANKING OF BRANCHES FOR IEEE 14-BUS SYSTEM Line Number Number of Over Loaded Lines (NOLL) Number of Voltage Violation Buses (NVVB) Line To Bus From Bus Performance Index PI = ( NOLL + NVVB )

Voltage distribution for IEEE 14-bus system by both GA and PSO techniques when line 3 is outage is shown in Fig. 7. Before placing UPFC in this system, there were 6 buses which have voltage violations, and after placing UPFC all of them are eliminated. These results are achieved by applying both GA and PSO techniques.
1.2 1.1 1 0.9 0.8 ) . u . 0.7 p ( 0.6 e g a t 0.5 l o V0.4 0.3 0.2 0.1 0 1 2 3 4 5 6 7 8 Buses 9 10 11 12 W/O UPFC GA PSO 13 14

1 3 2 10 4 14 15 13 5 7 8 9 16 17 20

1 2 1 5 2 7 7 6 2 5 4 4 9 9 13

2 3 5 6 4 8 9 13 5 4 7 9 10 14 14

7 4 2 1 3 0 0 1 1 0 0 0 0 0 0

6 5 6 5 2 3 3 2 1 1 1 1 1 1 1

13 9 8 6 5 3 3 3 2 1 1 1 1 1 1

Rank 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Fig. 7. Voltage distribution for IEEE 14-bus system when line 3 is outage by GA and PSO

Objective Function

From contingency analysis and ranking process performed on this system as shown in Table VI, we can find that there are fifteen lines which are the severest contingencies scenarios in this system. In the case of one of this lines is outage, the most overloaded lines and bus voltage violations will be encountered. Therefore, to investigate the effectiveness of UPFC on the system under such cases, we install UPFC in the network and apply both GA and PSO techniques to find the optimal location and the optimal parameter setting of UPFC which eliminate or minimize the overloaded lines and bus voltage violations. Fig. 6 shows the power flow distribution in the IEEE 14-bus system when line 3 is outage. Before placing UPFC in this system, there were 4 overloaded lines, and after placing UPFC three of them are eliminated and the power flow distribution for the fourth overloaded line is reduced.
2 1.8 1.6 1.4 x a m P / P 1.2 1 0.8 0.6 0.4 0.2 0 1 2 4 5 6 7 8 9 10 11 12 13 15 16 17 18 19 20 Line Number W/O UPFC GA PSO

The minimization of the objective function achieved by both GA and PSO techniques when line 5 is outage are shown in Fig.8 .Because of the limited space for this paper, we only choose this case as a sample of the results weve obtained.

9.40 9.30 9.20 9.10 9.00 8.90 8.80 8.70 8.60 8.50 GA PSO

10

20

30

40

50 60 Generation

70

80

90

100

Fig. 8. Minimization of the objective function by both GA and PSO techniques for IEEE 14-bus system when line 5 is outage.

Fig. 6. Power flow distribution for IEEE 14-bus system when line 3 is outage by GA and PSO

Table VII and Table VIII show the overloaded lines and bus voltage violations before and after placing UPFC in the IEEE 14-bus system with optimal location and optimal parameters setting of UPFC. These results are achieved by applying both GA and PSO techniques.

TABLE VII OVERLOADED LINES AND BUS VOLTAGE VIOLATIONS BEFORE AND AFTER PLACING UPFC FOR IEEE 14-BUS SYSTEM WITH OPTIMAL LOCATION AND OPTIMAL PARAMETERS SETTING OF UPFC BY GA Optimal Placement of UPFC After Placing UPFC Voltage Violation Buses Optimal Setting of UPFC Tripped Before Placing UPFC Line Voltage Violation Buses

settings achieved by applying both GA and PSO.

TABLE VIII OVERLOADED LINES AND BUS VOLTAGE VIOLATIONS BEFORE AND AFTER PLACING UPFC FOR IEEE 14-BUS SYSTEM WITH OPTIMAL LOCATION AND OPTIMAL PARAMETERS SETTING OF UPFC BY PSO Optimal Placement of UPFC Tripped Before Placing UPFC Line Overloaded Lines Voltage Violation Buses Overloading % After Placing UPFC Voltage Violation Buses Overloaded Lines Overloading % Optimal Setting of UPFC

Overloaded Lines

Overloaded Lines

Overloading %

Overloading %

Rank

From Bus

From Bus

To Bus

To Bus

VcR

VvR

VcR

VvR
1

1-5 2-3 2 3-4 4-5 9-10

110.91 191.96 189.11 170.62 133.14

1-5 2-5 3-4 4-5

119.38 135.28 190.52 175.41

1-2 2-5

142.96 163.71

4-5

157.28

Bus_2 Bus_4 Bus_5 Bus_6 Bus_7 Bus_13 Bus_14 Bus_3 Bus_9 Bus_10 Bus_12 Bus_13 Bus_14 Bus_9 Bus_10 Bus_11 Bus_12 Bus_13 Bus_14 Bus_6 Bus_11 Bus_12 Bus_13 Bus_14

2-3 174.49 3-4 138.21 4-5 144.24 -

6 0.1834 0.9001 1

1-5 2-3 2 3-4 4-5 9-10

110.91 191.96 189.11 170.62 133.14

3-4 171.82 -

2 6 0.0898 1.0995 2

1-5 2-5 3-4 4-5

119.38 135.28 190.52 175.41

1 4-5 106.46 7 0.1063 1.0992 3

1-2 2-5

142.96 163.71

5 4-5 125.02 19 0.0445 1.0245 4 2

4-5

157.28

Bus_2 Bus_4 Bus_5 Bus_6 Bus_7 Bus_13 Bus_14 Bus_3 Bus_9 Bus_10 Bus_12 Bus_13 Bus_14 Bus_9 Bus_10 Bus_11 Bus_12 Bus_13 Bus_14 Bus_6 Bus_11 Bus_12 Bus_13 Bus_14

1-5 108.65 2-4 112.33 4-5 138.98

6 0.1794

0.9 1

3-4 164.58 -

6 0.0898 1.0948 2

7 0.1076 1.0869 3

4-5 114.88

13 0.0299 1.0590 4

116.82 Bus_10 135.75 2 4 3-4 117.25 Bus_14 173.71 Bus_9 Bus_10 7 8 Bus_14 Bus_9 Bus_10 7 9 Bus_14 Bus_13 6 13 12-13 129.68 Bus_14 2 5 1-5 119.17 Bus_14 5 4 Bus_14 4 7 Bus_14 4 9 Bus_14 9 10 Bus_10 9 14 Bus_14 13 14 Bus_14 -

1-5 2-5 4-5

6 0.1217 0.9636 5

6 0.1059 0.9008 6

17 0.0279 1.0999 7 20 0.0525 0.9166 8 20 20 20 6 6 6 6 0.0462 0.0527 0.0513 0.0989 0.0984 0.0929 0.1058 0.9059 0.9003 0.9009 0.9002 0.9 0.9003 0.9007 9 10 11 12 13 14 15

116.82 Bus_10 135.75 Bus_14 173.71 4-5 132.26 Bus_9 Bus_10 7 8 Bus_14 Bus_9 Bus_10 7 9 Bus_14 Bus_13 6 13 12-13 129.68 Bus_14 2 5 1-5 119.17 Bus_14 5 4 Bus_14 4 7 Bus_14 4 9 Bus_14 9 10 Bus_10 9 14 Bus_14 13 14 Bus_14 4 1-5 2-5 4-5

6 0.1214 0.9623 5

6 0.1198 1.0822 6

17 0.0288

1.1 7

20 0.0453 0.9278 8 6 6 6 20 6 6 6 0.0908 0.1091 0.0915 0.046 0.1006 0.0793 0.1042 1.0121 0.9 0.9 0.9500 0.9009 0.9 0.9924 9 10 11 12 13 14 15

From Tables VII and VIII, we find that despite of placing UPFC in the system didnt result in eliminating all of the overloaded lines, but most of overloaded lines are eliminated and the power flow distribution in the rest overloaded lines is significantly reduced. While all of the bus voltage violations are eliminated by placing UPFC in an optimal location with optimal parameter

From the simulation results, we found that the applied GA and PSO techniques are suitable and efficient for the considered optimization problem. In the following we summarize the main differences that we have noticed from the implementation results: a) From the convergence perspective, it is observed that PSO technique is faster than GA technique in the beginning of the optimization, but with the increase of generation, it can be seen that the performance of GA is better than the

Rank

performance of PSO as shown in Fig. 5 and Fig.8. b) From the perspective of time, it was found that the PSO technique is much faster than the GA technique and this is due to the fact that GA technique has selection, crossover, and mutation operations, while PSO doesnt have such operations. V. CONCLUSION In this paper, the effectiveness of UPFC for enhancing the security of power systems under single contingencies has been investigated. Determination of the severest contingencies scenarios were done based on the contingency selection and ranking process. Two of the most powerful evolutionary optimization techniques namely: Genetic Algorithm and Particle Swarm Optimization have been successfully applied to the problem of optimal location and parameters setting of UPFC. Maximization of power system security was considered as an optimization criterion. Also, in this paper, two case studies were conducted using an IEEE 6-bus system and an IEEE 14-bus system. The obtained results show that both GA and PSO techniques have performed well and have superior features including high-quality solution, stable convergence characteristic, and good computation efficiency. Finally, our results show that the UPFC can significantly improve the security of power systems under single contingencies with optimal location and optimal parameter settings. VI. REFERENCES
N. G. Hingorani, Power electronics in electrical utilities: role of power electronics in future power systems, Proceedings of the IEEE, vol. 76 No. 4, pp.481-482, Apr.1988. [2] L. Gyugyi, A unified power flow control concept for flexible AC transmission systems, IEE Proc., Part-C, vol.139, No.4, pp.323-331, Jul. 1992. [3] C. R. Puerle-Esquivel, and E. Acha, A Newton-type algorithm for the control of power flow in electrical power networks, IEEE Trans. Power System, vol. 12, No. 4, pp. 1474-1480, Nov. 1997. [4] S. N. Singh, I. Erlich, Locating unified power flow controller for enhancing power system loadability, International Conference on Future Power System, pp.1 5, Nov. 2005. [5] W. Ongskul, and P. Jirapong, Optimal allocation of FACTS devices to enhance total transfer capability using evolutionary programming, IEEE International Symposium on Circuits and Systems, vol.5, pp.4175-4178, May. 2005. [6] K. P. Wang, J. Yurevich, and A. Li, Evolutionary-programming-based load flow algorithm for systems containing unified power flow controllers, IEE Proc.-Gener. Transm .Distrib, vol.150, No. 4, Jul. 2003. [7] L. Ippolito, and P. Siano, Selection of optimal number and location of thyristor-controlled phase shifters using genetic based algorithms, IEE proc.-Gener. Trasm. Distrib., vol. 151, No. 5, Sept. 2004. [8] D. Radu, etal, A multi-objective genetic algorithm approach to optimal allocation of multi-type FACTS devices for power system security, IEEE Power Engineering Society General Meeting, pp. 8, Jun. 2006. [9] M. Saravanan, etal, Application of PSO technique for optimal location of FACST devices considering system loadability and cost of installation, Power Engineering Conference, vol. 2 pp.716 -721, Dec. 2005. [10] H. I. Shaheen, G. I. Rashed, and S. J. Cheng, Optimal location and parameters setting of Unified Power Flow Controller based on evolutionary optimization techniques, IEEE PES General Meeting, Tampa, FL, USA, pp.8, Jun. 2007 [11] D .Sobajic, and Y. Pao, An artificial intelligence system for power system contingency screening, IEEE Trans. Power Systems, vol.3, pp 647-653, May 1988. [1]

[12] C. Ejebe, and etal., Fast contingency screening and evaluation for voltage stability analysis, IEEE Trans. Power System, vol.3, pp 1582-1588, Nov. 1988. [13] R .Chen, and etal., Multi-contingency preprocessing for security analysis using physical concepts and CQR with classifications, IEEE Trans. Power Systems, vol.8, pp 840-846, Aug. 1993. [14] S .Weerasooriya, and etal., Towards static security assessment of a large scale power system using neural networks, IEE Proceedings-C, vol. 139, pp 64-70, Jan.1992. [15] S. H. Song, J. U. Lim, S. W. Jung, and S. I. Moon, Preventive and corrective operation of FACTS devices to cope with a single line-faulted contingency, Power Engineering Society General Meeting, IEEE, vol.1, pp. 837 842, Jun. 2004. [16] A. Sudersan, M. Abdelrahman, and G. Radman, Contingency selection and static security enhancement in power systems using heuristics-based genetic algorithms, Proceedings of the Thirty-Sixth Southeastern Symposium, pp: 556 560, 2004. [17] D. Radu, and Y. Besanger, Blackout Prevention by Optimal Insertion of FCACTS Devices in Power Systems, In proc. IEEE Future Power Systems Conf., Amsterdam, NL. 2005 [18] D. E. Goldberg, Genetic Algorithms in Search, Optimization and Machine Learning, Addison-Wesley, 1989. [19] D. A. Coley, An Introduction to genetic Algorithms for Scientists and Engineers, World Scientific Publishing Co., 1999. [20] E. Bonabeau, M. Doringo, and G. Theraulaz, Swarm Intelligence: From Natural to Artificial Systems, New York: Oxford University Press, Santa Fe Institute Studies in the Science of Complexity, 1999. [21] E. Bonabeau, and G. Theraulaz, Swarm Smarts, Scientific America, No.3, 2000, pp.71-79. [22] R. C. Eberhart, J. Kennedy, and Y. Shi, Swarm Intelligence, San Francisco: Morgan Kaufman Publishers, 2001. [23] Y. J. Liu, X. X. He, Modeling Identification of Power Plant Thermal Process Based on PSO Algorithm, American Control Conference, pp.4484-4489, Jun. 2005. [24] A. J. Wood and B. F. Woolenberg, Power Generation, Operation and Control, Wiley, 1996. [25] L. L. Freris, and A. M.Sasson, Investigation on the load flow problem, Proceeding of IEE, vol. 115, pp.1459-1470, 1968. [26] A. Enrique, and etal., FACTS Modeling and Simulation in Power Network, John Wiley &Sons, LTD. 2004

VII. BIOGRAPHIES
Husam I. Shaheen, student member IEEE, received his Bachelor degree and his Diploma in Electrical Engineering from Techreen University, Syria, in 1999 and 2000 respectively, and his Master degree in Power System Automation, from Huazhong University of Science and Technology (HUST), China in 200.5. Recently he is a PhD candidate at Huazhong University of Science and Technology (HUST). His main interest researches are AI and its application to power system, FACTS devices and their control. Ghamgeen I. Rashed, student member IEEE, was born in Sulaimani-Iraq, on Sept. 16, 1974. He received his B.Sc. in Electrical Engineering in Salahaadin University-Iraq at 1995, and his M.Sc. in Sulaimani University-Iraq in 2003.Now he is a ph.D candidate in Huazhong University of Science and Technology (HUST), China. His special research of interest is AI and its application to power system, FACTS devices specially TCSC and its control. Shijie Cheng, senior member, IEEE, Got his Ph.D. degree in Canada in 1986. He is a life professor of Huazhong University of Science and Technology, China. In recent years he has been engaged in the area of powerline communication, intelligent control, stabilization control of power system and superconductivity power technology.