Effect of Multi-SVC Installation for Loss Control in Power System using Multi-Computational Techniques

— Flexible AC Transmission Systems (FACTs) play a vital role in minimizing the power losses and improving voltage profile in power transmission system. These increase the real power transfer capacity of the system. However, optimal location of sizing of the FACTs devices determines the extent of benefits provided by the FACTs devices to the transmission system. Non-optimal solution in terms of the location and sizing may possibly lead to under-compensation or over-compensation phenomena. Thus, a robust optimization is a priori for optimal solution achievement. This paper presents a study on the effect on multi static VAR compensators (SVC) installation for loss control in power system using evolutionary programming (EP), artificial immune system (AIS) and immune evolutionary programming (IEP). The objective is to minimize the real power loss transmission and improve the voltage profile of the transmission power system. The study reveals that installation of multi-units SVC significantly reduces the power loss and increases the voltage profile of the system, validated on the IEEE 30-Bus Reliability Test System (RTS).


INTRODUCTION
The world electricity demand has been steadily increasing over the past few decades due to various factors such as population growth, urbanization, economic development, and technological advancements [1]. Conventional approach such as construction of new power plant and transmission lines to meet the increasing demand of electricity is not a feasible option due to many reasons such as high cost, environmental concerns, technical and time constraints. One of the vital ways to optimize the current system is by reducing the loss in the power system which can be achieved by injecting or retrieving the reactive power with the assistance of FACTs. Installation of FACTs devices can be one of the popular initiatives.
The concept of FACTs was first introduced by Hingorani in 1998 [2]. The working principle of FACTS is to enable the system electric parameters to change fast and flexibly while maintaining the security, stability and reliability of power system; thus, optimizing the existing resources in reducing power loss and cost, while improving the efficiency of the power grid operation [3]. It is important to understand that FACTs devices confined several compensation devices such as unified power flow controller (UPFC), static VAR compensator (SVC), thyristor-controlled series compensator (TCSC), and static synchronous compensator (STATCOM). SVC is considered as a popular compensating device in power system. SVC is a parallel connected device which can act as variable capacitor or variable inductor.
The important uses of SVC are voltage stabilization of weak networks, reducing transmission losses, increasing power transfer capacity, increasing the small disturbance damping, improving voltage stability and removing power fluctuations [4]. SVC is chosen in this work based on its functionality. Optimal placement and sizing remain as a challenge in application of SVCs in voltage stability enhancement [5], [6]. Several optimization techniques have been invented to achieve the optimal solution for SVC installation for the purpose of minimizing the loss in power system or controlling the voltage level in a system within the acceptable limit determined by IEEE or IET standards. Otherwise, a power system will operate under low voltage level which in turn affecting the life time of a transmission cable in the system. Heuristic optimization techniques have become an important approach in solving complex which is difficult to be solved using traditional approach [6], [7]. Heuristic optimization technique has been adopted in determining the optimization of FACTs devices for power system performance improvement. EP was used to optimize the size of SVC for minimization of loss and voltage profile improvement [8]. EP and AIS were used for SVC location and sizing optimization in IEEE 30-Bus RTS [9]. This application indicates that both EP and AIS is robust in achieving the optimal solution for SVC installation in power system.
On the other hand, improved version of traditional optimization techniques has also been proposed. Among the important techniques that can be highlighted, is a novel improved differential evolutionary (IDE) algorithm which is applied to optimize SVC and TCSC location and sizing for reactive power management in IEEE Bus-30 RTS, IEEE-57 RTS and IEEE-118 RTS [10]. Improved particle swarm in [11] is successfully applied in solving multi-objectives problems. However, the application dealt with the available transfer capability (ATC) study. Other important work that can be highlighted is the work conducted [12] where SVC and www.ijacsa.thesai.org TCSC were used as the compensating devices. Gravitational search algorithm (GSA) is proposed for the loadability enhancement of the power system under different loading conditions by determining the optimal placement of different TCSC and SVC [13]. Another work conducted using whale optimization technique in [14] can be considered useful involving SVC and TCSC installation. Thus, a critical review may lead to an efficient decision in any compensation effort in power system. This paper presents effects of multi-SVC installation for loss control in power system using multi-computational techniques. In this study, EP, AIS and Immune EP (IEP) were applied to identify the locations and sizing of SVC installation in controlling the transmission loss in power system. Controlling transmission loss in power systems is important as it enables optimal voltage control, reactive power compensation, and power flow regulation. By strategically placing SVCs, voltage stability is maintained, transmission losses are minimized, and system efficiency is improved, leading to reliable and secure operation. Results validated on the IEEE 30-Bus RTS revealed that all the techniques are comparable for loss control scheme in power system. This paper is organized as follows. The objectives and problem formulations are established in Section II. Section III describes the optimization techniques used in detail. Section IV presents the test system used in the work. Section V presents the simulation results and discussions. Finally, Section VI concludes the paper with recommendations.

II. OBJECTIVE AND PROBLEM FORMULATION
Loss minimization is a vital remedial action applied in power system planning and operation. Transmission loss is proportionally translated to monetary loss as not all electricity generated is delivered to the customers. Compensation scheme such as installation of SVC can greatly reduce losses in the system. This research holds significant importance as its main objective is to study the effect of installing multi SVCs for loss control in power system using multi-computational intelligence techniques. The optimization of SVC location and sizing plays a crucial role in achieving this objective. By formulating the problem in terms of power loss minimization, this study addresses a critical aspect of enhancing power system efficiency and sustainability, thereby contributing to the advancement of power grid operations and control strategies.
The objective function, (OF) of this work is to minimize total loss in a transmission power system, which is mathematically represented by: where is number of buses in the system, and is power loss for line which can be determined using (2), (3) and (4). The objective function is subjected to voltage constraints and SVC location and sizing during the optimization process. The minimum voltage must be kept within the specified limits during the optimization process as indicated in (5). The sizing of the SVC is subjected to constraint as shown in (6).

III. OPTIMIZATION TECHNIQUES
Optimization technique is a collection of mathematical principles and methods used in solving a quantitative problem. Optimization technique is one of the most preferred and widely used techniques to solve SVC location and sizing optimization for minimization of transmission loss. In this study, three promising and evolving optimization techniques are applied, namely EP, AIS, IEP for location and sizing optimization of SVC for power loss reduction.

A. Evolutionary Programming
Evolutionary Programming (EP) is an artificial intelligence method inspired from natural selection process to find the global optimum of a complex problem. It was first invented by Lawrence J. Fogel in the US in 1960 [15]. The mechanics of EP algorithm is illustrated in Fig. 1.
Step 1: Initialization The initialization process of EP begins with generating the control variables for optimal location and sizing of SVC using a uniformly distributed random number generator. 20 www.ijacsa.thesai.org individuals (parents) are generated based on the control variables for the first iteration. In this case, the control variables will represent the locations and sizing of the SVC units. Apparently, the number of control variables or decision variables will be doubled of the number of SVC units to be installed into the system. During initialization, parameters such as: number of individuals, mutation step size and maximum number of iterations are specified. The total loss of the system before optimization, is calculated as the reference value. The individuals that violate the requirements and constraints are subsequently exterminated from the population pool. Step 2: Fitness 1 Calculation Load flow programs are performed to calculate the fitness value which is the real power loss, . If and fulfils the other constraints specified, they will be accepted into the initial population pool. The general matrix for the initial population, is given by (9).
[ ] where: = population size = no. of variable = fitness of the individual (parent) Step 3: Mutation Mutation is a process to breed offspring. The individuals (parents) in the population pool are mutated to breed offspring using the Gaussian mutation operator [16]- [18]. The Gaussian mutation equation is given in (10). Step 4: Fitness 2 Calculation Fitness 2 calculation is performed by running the load flow again, using the offsprings bred during the mutation process. The matrix for the offspring population, is given by (11).
[ ] where = population size = no. of variable = fitness of the offspring Step 5: Combination Process The parent matrix and the offspring matrix are then combined in a cascaded form as in (12).
Step 6: Selection Process and Convergence Test The individuals in matrix is ranked in ascending order based on fitness value. 20 best individuals from matrix are selected for the next iteration. New individuals are identified for the next iteration process. In this case, the new individuals will be equipped together with the corresponding fitness value for the next iteration or evolution. However, if the new individuals do not bring together the corresponding fitness values, Fitness 1 calculation needs to be conducted. A convergence test is used to check if the optimal solution has been achieved. Otherwise, Step 2 through Step 4 www.ijacsa.thesai.org will be repeated until the convergence criterion is met or the iteration cycle reaches the maximum value set. EP algorithm converges when the difference between the maximum and minimum fitness values is 0.00001 presented mathematically in (13). (13) That ends the EP process.

B. Artificial Immune System
Artificial Immune System (AIS) technique is inspired by biological immune system and has been used for computational models in solving complex real-world problems. The algorithms in artificial immune system adapt the immune system features of learning and memory to solve a problem. The evolution of AIS has its roots from the work of Farmer, Packard and Perelson [19]. The AIS flowchart is presented in Fig. 2. Step 1: Initialization The initialization process of AIS begins with generating the control variables for optimal SVC using a uniformly distributed random number generator. 20 individuals (parents) are initially generated based on the control variables for the first iteration. During initialization, parameters such as: number of individuals, mutation step size and maximum number of iterations were to be specified. The total loss of the system before optimization, is calculated as the reference value. The individuals that violate the requirements and constraints are subsequently exterminated from the population.
Step 2: Parent Fitness Calculation Load flow programs is performed to calculate the fitness value which is real power loss, . If and fulfils the other constraints specified, they are accepted into the initial population pool. The general matrix for the initial population, is given by (9).
Step 3: Cloning Process The individuals in the initial pool, which are referred as parents, are cloned with a factor of 10 as suggested [20], [21]. The population size increases 10 folds after cloning. The general cloned matrix is given in (14) where is the cloning factor.
Step 4: Mutation Mutation is a process to produce offspring. The individuals of the cloned population, are mutated to breed offspring. Mutation equation in (10) is adapted.
Step 5: Offspring Fitness Calculation Then, load flow analysis is performed to determine the fitness of the offspring. The matrix for the offspring population, is given by (15). Step 6: Selection Process and Convergence Test Selection process is conducted to identify the survivors among the fittest. In this case, the individuals in matrix are ranked in ascending order based on fitness value. For this study, the individuals are ranked based on the lowest fitness value since the objective function is www.ijacsa.thesai.org minimization of total transmission loss. 20 best individuals from the matrix are selected for the next iteration. A convergence test is used to check if the optimal solution has been achieved. Otherwise, Step 2 through Step 5 will be repeated until the convergence criterion is met or the iteration cycle reaches the maximum value set. AIS algorithm converges when the difference between the maximum and minimum fitness values is 0.00001as presented in (13). That ends the AIS process.

C. Immune Evolutionary Programming
Immune Evolutionary Programming (IEP) is derived by integrating the traditional EP and AIS. The IEP flowchart is presented in Fig. 3. Step 1: Initialization The initialization process of EP begins with generating the control variables for optimal SVC using a uniformly distributed random number generator. 20 individuals (parents) are generated based on the control variables first iteration. During initialization, parameters such as: number of individuals, mutation step size and maximum number of iterations were to be specified. The total loss of the system before optimization, is calculated as the reference value. The individuals that violate the requirements and constraints are subsequently exterminated from the population.
Step 2: Fitness 1 Calculation Load flow programs are run to calculate the fitness value which is real power loss, . If and fulfils the other constraints specified, they will be accepted into the initial population pool. The general matrix for the initial population, is given by (9).
Step 3: Cloning The individuals in the initial pool which are referred as parents are cloned with a factor of as in (14).
Step 4: Mutation Mutation is a process to produce offspring. The individuals in the cloned population pool are mutated to breed offspring. The Gaussian mutation equation in (10) is adapted.
Step 5: Fitness 2 Calculation Then, load flow analysis is performed to determine the fitness of the offspring. The matrix for the offspring population, is obtained as shown in (15).
Step 6: Combination Process The parent matrix and the offspring matrix are then combined in a cascaded form. If the parent matrix and the offspring matrix are represented by (9) and (15), respectively, then the combined matrix, C, has the form as in (16).
Step 7: Selection Process and Convergence Test The individuals in matrix are ranked in ascending order based on fitness value. 20 best individuals from matrix are selected for the next iteration. A convergence test is used to check if the optimal solution has been achieved. Otherwise, Step 2 through Step 4 will be repeated until the convergence criterion is met or the iteration cycle reaches the maximum value set. IEP algorithm converges when the difference between the maximum and minimum fitness values is 0.00001as presented in (13).
That ends the IEP process.

IV. TEST SYSTEM
The single line diagram of IEEE 30-Bus RTS used in this study [22] is shown Fig. 4, depicts the configuration of the power system. This system consists of five generators located at Bus 2, Bus 5, Bus 8, Bus 11, and Bus 13. There is a slack bus located at Bus 1 which serves as the reference point for voltage and frequency. There are 24 load buses in the system. The system is interconnected through a total of 41 bus interconnecting lines, forming a complex network. This power system topology is used for identifying and analyzing optimum location and sizing of SVC installation in controlling the transmission loss in power system. www.ijacsa.thesai.org

V. RESULTS AND DISCUSSIONS
This section presents the results of multi-SVC installation for loss control in power system using multi-computational techniques.

A. Bus Category Identification
Test was performed on IEEE 30-Bus RTS to identify the weak and strong buses of this system. An increasing reactive power load, Q d was subjected to each of the load bus in the system and the corresponding voltage at the bus is recorded until the system collapsed. The results are presented in Fig. 1.  Table I.

B. Multi SVCs Installation
In this study, EP, AIS and IEP are validated on IEEE 30-Bus RTS to determine the effect of SVC installation on power loss and voltage profile of the system. In this study, three scenarios have been considered as follows:  Scenario 1: Single SVC installation. Error! Reference source not found.TABLE 2 tabulates the implementation scope of the multi-SVC installation on IEEE 30-Bus RTS. These are the scenarios studied to evaluate the proposed optimization algorithm in finding the optimal location and sizing of SVCs.

1) Scenario 1: Single SVC installation:
This section discusses the optimization results of single SVC installation in the attempt to minimize the of the system. All together 6 buses were subjected to reactive power load; 3 weak buses (Bus 26, 29, 30) and 3 strong buses (Bus 6, 4, 3).
III presents the results for single SVC installation involving weak buses at Buses 26, 29 and 30 and strong buses involving Buses 6, 4 and 3. In general, all techniques managed to reduce the power loss in the system. For instance, when Bus 26 was subjected to 10 MVAR, the installation of single SVC managed to reduce the loss from 18.230 MW to 17.820 MW, solved using EP. The LRP is 2.249%. When the load is increased to 30 MVar, loss was reduced from 26.109 MW to 17.542, where the LRP is 32.815%. AIS and IEP also managed to achieve these results, which imply that all the three techniques are comparable. Generally, the values of LRP increase as the reactive power loading is increased regardless of any optimization technique. The details can be referred to the same table.
Table IV presents V min and VIP before and after the single SVC installation. When Bus 26 was subjected to 10 MVAR, the installation of single SVC managed to increase the V min of IEEE 30-Bus RTS from 0.940 p.u. to 0.957 p.u., solved using EP. The VIP is 1.841%.
When the load is increased to 30 MVAR, the V min increased from 0.691 p.u. to 0.973 p.u., where the VIP is 40.799%. AIS and IEP also managed to achieve these results. Generally, the values of VIP increase as the reactive power loading is increased regardless of any optimization technique. More details can be referred to the same table.    2) Scenario 2: 2 SVCs installation: Table V presents the results for 2 SVCs installation involving weak buses and strong buses. In general, all techniques managed to reduce the power loss in the system with 2 SVCs installation. For instance, when Bus 26 was subjected to 30 MVAR, the installation of 2 SVCs managed to reduce the loss from 26.109 MW to 17.663 MW, solved using EP. The LRP is 32.352%. AIS and IEP also managed to achieve similar results. On the other hand, when strong buses (Bus 6, 4, 3) were subjected to 50 MVAR, the installation of 2SVCs managed to reduce the losses, solved using all the three techniques. However, the LRP in the range of 2.5% to 5% only. This implies that SVC installation reduces the loss in the power system greatly when weak buses are subjected to reactive power load compared to strong buses.
Tale VI presents V min and VIP before and after the 2 SVCs installation. Overall, 2 SVCs installation increases the V min in both weak buses and strong buses regardless the optimization techniques applied. For instance, when Bus 30 is subjected to 30 MVAR, V min is increased from 0.707 p.u. to 1.010 p.u., solved using EP. The VIP is 42.877%. AIS and IEP also managed to achieve VIP in the similar range. Details on other buses can be referred to Table VI. Table VII presents the results for 3 SVCs installation involving weak buses and strong buses. 3 SVCs installation reduced loss in the system regardless the techniques applied. When Bus 30 is subjected to 30 MVAR, the loss was reduced from 26.109 MW to 19.140 MW which implies a LRP of 26.695% solving with EP. For the same scenario the LRP was 32.815% for single SVC installation and 32.352% for 2 SVCs installation. Similar results were observed with AIS and IEP. These results indicate, installation of multi-SVC not necessarily further reduces the loss of the power system. Based on the results, a maximum of 2 SVCs installation is recommended. However, other constraints of the power system need to considered. The details can be referred Table VII.   Table VIII presents V min and VIP before and after the 3 SVCs installation. Overall, 3 SVCs installation increases the V min in both weak buses and strong buses regardless of the optimization techniques applied. The best VIP was recorded when Bus 30 was subjected to 30 MVAR. The V min was increased from 0.691 p.u. to 1.010 p.u. solving using EP. The VIP is 46.228%. AIS and IEP also managed to achieve same results. Results for other buses can be referred to

C. Location and Sizing of SVCs
The installation of SVCs into the system as the initiative to reduce the total power loss will require the optimal locations and sizing. The multi-SVC locations and sizing details for all the three scenarios are tabulated in Tables IX to XI. However, it was observed the sizing of single SVC is much lower for the same scenario compared to total SVC sizing when 2 SVCs installation was opted. In this case, single SVC and 2 SVCs installations has same range of LRP. This implies, multi-SVC is not always an option to reduce loss in the system. Other factors such as installation cost, maintenance cost and accessibility need to be considered before opting for multi-SVC installation. Details of 2 SVCs location and sizing when other buses are subjected to Q d,max can be referred to Table X On the other hand, it was observed the sizing of single SVC and 2 SVCs are much lower for the same scenario compared to total SVC sizing when 3 SVCs installation was opted. In this case, single SVC and 2 SVCs installations have higher LRP compared to 3 SVCs. This implies, multi-SVC is not necessarily an option to obtain the highest LRP in the system. Based on the results, installation of 3 SVCs in not recommended as the LRP achieved is lower compared to single SVC and 2 SVCs. Other factors such as installation cost, maintenance cost and accessibility need to be considered before opting for multi-SVC installation. Details of 3 SVCs location and sizing when other buses are subjected to Q d,max can be referred to Table XI. VI. CONCLUSION Based on the key findings of this study, it can be concluded that the three optimization techniques, namely EP, AIS and IEP are effective solutions for addressing the complex task of optimal location and sizing of multi-SVC installation in the IEEE 30-Bus RTS power system. The work successfully explored the optimal positions, sizes, and number of SVCs while considering network operational constraints and SVC limitations, ultimately aiming to reduce total active power loss.
Findings of this study have broader implications for the field of power system optimization and control. It contributes to advancing the understanding of SVC installation for enhancing power system performance using multicomputational intelligence techniques It also showcases the effectiveness of EP, AIS and IEP in tackling large-scale technical challenges and highlights their potential for addressing similar optimization problems in other power system scenarios.
In summary, this work demonstrates the efficacy of EP, AIS, and IEP in solving the optimal location and sizing problem of multi-SVC allocation. These results provide valuable insights for power system optimization and have broader implications for improving power grid performance, reliability, and efficiency.