Self-Adaptive Firefly-Algorithm-Based Unified Power Flow Controller Placement with Single Objectives

(e selection of positions for unified power flow controller (UPFC) placement in transmission network is an essential factor, which aids in operating the system in a more reliable and secured manner. (is paper focuses on strengthening the power system performance through UPFC placement employing self-adaptive firefly algorithm (SAFA), which selects the best positions along with parameters for UPFC placement. (ree single objectives of real power loss reduction, voltage profile improvement, and voltage stability enhancement are considered in this work. IEEE 14, 30, and 57 test systems are selected to accomplish the simulations and to reveal the efficacy of the proposed SAFA approach; besides, solutions are compared with two other algorithms solutions of honey bee algorithm (HBA) and bacterial foraging algorithm (BFA). (e proposed SAFA contributes real power loss reduction, voltage profile improvement, and voltage stability enhancement by optimally choosing the placement for UPFC.


Introduction
Power system stability is the most challenging for power system engineers due to the ever increasing load demand; as a result, the power system network falls to stressed condition, which leads to loss reduction and voltage instability. Moreover, it is a challenge to widen the existing network to satisfy the growing power demand by setting up new generation and transmission networks owing to economic and environmental constraints. Installation of flexible alternating transmission system (FACTs) controllers in transmission network is a possible solution to overcome the stability issues.
FACTs controller is classified into series, shunt and blending of series and shunt controller. Static var compensator (SVC) belongs to shunt controller, thyristor controlled series compensator (TCSC) comes under series controller, and UPFC associates with a combination of series and shunt controller. UPFC has the capability to serve as a series and shunt controller simultaneously; thereby, it controls the power flow, bus voltages, and phase angle with altering transmission line impedance [1][2][3].
Power system researchers have been paying their attention to carry out their research on optimal FACTs placement in transmission network for the last three decades to strengthen the existing power system performance. Nonconventional optimization algorithms of GA, DE, PSO, Honey Bees Algorithm (HBA), Ant Colony Optimization (ACO), Bacterial Foraging Algorithm (BFA), Firefly Algorithm (FA), Whale Optimization Algorithm (WOA), and Gravitational Search Algorithm (GSA) are commonly employed to solve power system optimization problems as they offer high-quality solutions .
Power system available transfer capability has been improved by the FACTs placement using HBA [4]. Evolutionary algorithms have been proposed to select the feasible locations and parameters of TCSC for increase of the power flow, reduction of losses, and enhancement of stability of the system, and also performances are tested in IEEE 14 bus system [5]. A fuzzy lag-lead controller has been used to control the parameters of TCSC and SVC for oscillation damping, stability enhancement, and ant colony optimization algorithm applied for setting of parameters of TCSC and SVC [6].
Particle swarm optimization was employed to improve power transfer capability and economic power system operation through proper SVC and TCSC placement [7]. A complete review of application of particle swarm optimization for FACTs placement in power system network has been reported [8]. A new Honey Bee Algorithm (HBA) has been proposed to overcome the limitation of conventional HBA, which offers near optimum solution for economic dispatch problem. e proposed algorithm improves matting process by combining chaotic local search, which has been applied in order to solve economic dispatch problem [9].
Bacterial foraging algorithm (BFA) has been applied to solve optimization problem more than it has been employed for minimizing loss and enhancing voltage profile through UPFC placement [10]. BFA has been employed for capacitor placement in radial distribution system in which BFA determines the optimal locations and size for capacitor placement in order to improve the performance of the system in terms of reducing power loss and improving voltage profile [11]. Dr. Yang developed firefly algorithm (FA) for solving optimization problem [12]. e FA has been attracting power system researchers for solving power system optimization problems of economic dispatch and unit commitment [13,14]. e limitation of FA includes slow process, lower convergence rate, and offering suboptimal solution. In order to overcome the limitations of FA, SAFA has been applied for placement of SVC, UPFC, and TCSC. e research work has addressed only one objective of loss minimization through SVC, UPFC, and TCSC placement [15][16][17]. Consequently, the author has considered multiobjectives for FACTs placement using SAFA in which SVC, TCSC, and UPFC have been chosen as FACTs devices for their placement to strengthen the power system performance by loss reduction, VP improvement, and VS enhancement [1]. e shunt-connected FACTs devices of SVC and STATCOM have been considered to compare their operating characteristics, and the solutions reveal that STAT-COM provides better advantages than SVC [18]. e whale optimization algorithm has been proposed for SVC and TCSC placement in IEEE 30 and 57 test systems. ey reported objectives of loss reduction and voltage profile improvement through shunt-and series-connected FACTs devices placement. In addition, WOA has been used for designing multi-input single-output controller of SSSC for stability enhancement [19,20].
An evolutionary algorithm of teaching-learning-based optimization was addressed for loss reduction, cost minimization, and voltage deviation reduction through TCSC placement. However, the author failed to address the way of selecting the number of TCSC for their placement [21]. A gravitational search algorithm has been presented for TCSC placement for controlling congestion management in deregulated environment and analysed by considering normal and contingency conditions [22].
Ant colony optimization has been proposed for tuning the gain parameters of PI controller, which has been used to control the gate signal of SSSC in order to damp oscillations and deviations in voltage of multimachine power system. Subsequently multimachine system stability has been enhanced by the application of ACO [23]. Artificial bee colony algorithm has been presented for solving reactive power dispatch problem aiming at loss minimization and voltage stability enhancement [24].
UPFC has been considered to improve the transmission capacity by optimal placement and the solutions are presented with comparisons [25]. An updated complete review has been presented for multitype FACTs devices placement and different metaheuristic techniques employed for their placement [26]. Artificial neural network has been applied for monitoring online voltage of power system. In addition voltage stability enhancement and loadability have been increased by optimally placing FACTs devices in IEEE 14 bus system [25][26][27][28][29][30][31][32][33][34].
In recent years, the FACTS devices attract the system engineers and researchers for providing better adaptation to varying operational conditions and improving the usage of existing installations. e placement of FACTS devices can be described as an optimization problem with the objective of minimizing the network loss. Based on the above findings, the researchers failed to concentrate on the placement of UPFC with different single objectives.
ere is thus a need for developing better strategies for optimally selecting the parameters with a view of obtaining the global best solution besides achieving better convergence. So, in this work, Self-Adaptive-FA-(SAFA-) based strategies have been proposed to minimize the transmission loss through placing TCSCs [14] and UPFCs effectively. us, our work provides the following: (i) Real power loss reduction (ii) Voltage profile improvement (iii) Voltage stability improvement (iv) A strategy of new SAFA employed to identify feasible positions for UPFC placement (v) ree IEEE 14, 30, and 57 bus systems considered (vi) Simulated results are compared with other two metaheuristic optimization algorithm solutions (vii) e self-adaptive scheme attempts to prevent suboptimal solution and enhance the convergence of the algorithm

Firefly Algorithm
It is a firefly-based metaheuristic optimization algorithm for solving power system optimization problem [12]. e light intensity of two fireflies' r and s decides the movement of attraction of two fireflies. Light intensity (LI) of rth firefly is represented by vector (X r ) as in 2 Complexity e attractiveness parameter between two fireflies of r and s is represented as β r,s � β max ,r,s − β min ,r,s exp −c r r 2 r,s . (2) e Cartesian distance between fireflies' r and s is as follows: Firefly position is being updated at the end of each iterative step by their movement towards brighter firefly. e movement of firefly "r" towards firefly "s" at m th iteration is presented in (4)

Self-Adaptive Firefly Algorithm (SAFA).
In SAFA, individual firefly decision variables include firefly parameters such as random movement factor (α), attractiveness parameter (β min ), and absorption coefficient (c). In previous studies, the researchers have used firefly algorithm for placement of FACTs devices. In firefly algorithm, the firefly parameters are fixed. e firefly parameters need to be tuned manually to obtain feasible solutions after each iteration. In this proposed SAFA, the parameters are tuned by selfadaptive mechanism in each iterative step by inclusion of three parameters namely α r , β min ,r , c with the number of decision variables of representation of firefly equation (1) and the firefly is represented as Each firefly with their parameters undergoes a whole search process; however, (2) is modified based on the brightness of fireflies as follows: e advantage of SAFA includes less computational effort, avoiding the suboptimal solution and convergence enhancement.

Mathematical Modelling of UPFC
Real and reactive power flow between the buses i and j are represented by [1] Real power, UPFC is modelled by combining series and shunt FACTs controllers. TCSC belongs to series controller, and its reactance (X tc ) is decided by the compensation factor and transmission line reactance (X line ) in which they are connected. e modelling of TCSC is formed by its reactance and is presented as SVC is one of the shunt controllers and is used to modify/control bus voltages through reactive power generation/absorption. SVC alter the reactive power (ΔQ i ) of the bus at which they are connected as follows: In order to model the UPFC, the conventional converters of the UPFC are replaced through voltage/current sources, which alter Jacobian elements based on active and reactive power injections in the buses. In this work SVC and TCSC are combined to make UPFC model using (8) and (9) to evade alterations in power flow and Jacobian structure.

Proposed Strategy/Problem Formulation.
e UPFC placement is chosen as an optimization problem and an objective is given by the following expression: x is the set of dependent variables consisting of active and reactive power of slack bus, reactive power of generator bus, and real power loss. y is the vector of independent variables comprising location for UPFC placement and their parameters. e equality constraint is the set of nonlinear power flow equations, which is presented as follows: P(V, δ) − P sp � 0, for generator and load buses, Inequality constraints are defined by the following expressions: In this work, three different single objectives (cases) are chosen through UPFC placement as follows.
Complexity Case 1. Real power loss (P loss ) e real power loss (P loss ) minimization is expressed as follows [1]: Case 2. Voltage profile improvement Total voltage deviation (TVD) is minimized to improve the bus voltage profile as per the following equation [1]: Case 3. Voltage stability enhancement e objective is formed to enhance voltage stability by minimizing L-index as per the following equation [1]: Values of F ij are obtained through bus admittance matrix. Table 1 gives the information about the SAFA variables. e first row of Table 1 represents the transmission line number for the placement of UPFC and the subsequent rows represent reactive power injection, Q f , compensation factor, and parameters of firefly.

Fitness Function.
e SAFA maximizes the fitness function of light intensity function in order to provide the optimal solution.

Simulation and Discussion
e simulation is performed on MATLAB software to analyse feasibility of SAFA for three-single-objective optimization through UPFC placement in standard IEEE 14, 30, and 57 test systems. Figure 1 demonstrates the flow of SAFA approach for UPFC placement.
Step-by-step procedure of SAFA for UPFC placement.
Step 1: read the considered system data.
Step 2: initialize the firefly parameters such as random movement factor (α), attractiveness parameter (β min ), absorption coefficient (c), number of fireflies, and maximum number of iteration.
Step 3: generate the initial population of fireflies and firefly position is initiated using (5).
Step 4: set iteration counter and run the simulation to calculate the brightness of all considered fireflies.
Step 5: firefly position is modified based on the brightness of firefly. e new position of firefly is obtained by using (6).
Step 6: obtain the firefly parameters from rth and sth firefly.
Step 7: place the UPFC devices according to the location and parameters available in the rth and sth fireflies.
Step 8: run the load flow and compute the performances and light intensity of rth and sth fireflies.
Step 9: the firefly with highest light intensity in the population is the optimal solution.
Initially the SAFA approach is run considering diverse number of UPFCs for P loss minimization to select the number of UPFC for placement; besides, the solutions are presented in Table 2. It is possible to observe from  Evaluate r r,s and β r,s by equations. (3) and (     In Case 2, an objective is chosen for the improvement of bus voltage profile by reducing total voltage deviation (TVD). It is possible to notice from Case 2 of Table 3 that TVD in 14 bus system reduced drastically from 0.3822 to 0.0603 by SAFA, whereas HBA and BFA decreased to 0.0657 and 0.0694, respectively. Likewise, TVD decreased dramatically from 0.4562 to 0.1640 via SAFA in 30 bus system (Table 4); however, HBA and BFA decreased to 0.1871 and 0.1974, respectively. In addition, TVD in 57 bus system reduced significantly from 1.2195 to 0.8194 by SAFA (Table 5), while HBA and BFA dropped off TVD to 0.9753 and 0.9854, respectively. In addition to this it can be seen from Case 2 of Tables 3-5 that all three algorithms provide solution of load bus voltage magnitudes to be nearer to one per unit as the chosen objective is voltage profile improvement.
us, the solutions reveal that SAFA selects the best positions for UPFC placement with suitable parameters to improve VP more than HBA and BFA approaches. e objective chosen in Case 3 is the enhancement of voltage stability (VS) by minimizing MVSI. It is possible to notice from Case 3 of Table 3 that MVSI in 14 systems is reduced substantially from 0.0751 to 0.0488 by SAFA, while HBA and BFA lessen to 0.0512 and 0.0538, respectively. Likewise, in 30 bus system (Table 4), MVSI considerable reduction is 0.07772 from 0.1420 by SAFA; however, HBA and BFA decrease to 0.0832 and 0.0844, respectively. Similarly, in 57 bus system (Table 5), MVSI is reduced significantly from 0.2914 to 0.2233 by SAFA, whereas HBA and BFA reduce MVSI to 0.2483 and 0.2683, respectively. In addition to this, it is possible to notice from Case 3 in Tables 2-5 that voltage magnitudes are resting within the range offered by algorithms. Since P loss and TVD minimizations are not considered as objective in this case, P loss and TVD are higher in three test systems. us, the solutions reveal that SAFA selects best positions for UPFC placement with suitable parameters to enhance VS more than HBA and BFA approaches. Figures 2-4 demonstrate the percentage of P loss savings rendered by optimization algorithms after UPFCs placement in 14, 30, and 57 bus systems, respectively. SAFA renders power loss saving as 1.16%, whereas HBA and BFA offer 1.08% and 0.97%, respectively, in 14 bus system. Similarly, in 30 bus system, P loss savings are 1.96%, 1.79%, and 1.78% by SAFA, HBA, and BFA, respectively. Likewise, in 57 bus system P loss savings are 3.35%, 0.75%, and 0.56% by SAFA, HBA, and BFA, respectively. Figures 2-4 demonstrate clearly that SAFA approach offers higher power loss savings than HBA and BFA approaches. Figures 5-7 illustrate the percentage of voltage profile (VP) improvement rendered by optimization algorithms after UPFCs placement in three test systems. In 14 bus system SAFA renders VP improvement as 84.22%, whereas HBA and BFA offer 82.81% and 81.84%, respectively. Similarly, in 30 bus system, VP improvements bestowed by SAFA, HBA, and BFA are 64.05%, 58.99%, and 56.73%, respectively. Likewise, in 57 bus system VP improvement offered by SAFA, HBA, and BFA is 32.81%, 20.02%, and 19.2%, respectively. It can be seen from Figures 5-8             identifies the best positions with appropriate parameters for UPFC placements to minimize P loss , improve VP, and enhance VS of the existing system. Tables 6-8 show information about the line locations and parameters for UPFC placement offered by SAFA. Table 9 provides information about control variables used for SAFA. One line diagram of IEEE 14,30, and 57 test systems are shown in Figures 17-19.
It is very clear from the above discussions that the proposed SAFA-based strategy is able to reduce the loss to the lowest possible value than those of other strategies. In addition the self-adaptive nature of the algorithm avoids        repeated runs for fixing the optimal FA parameters through a trial-and-error procedure and provides the best possible parameters values.

Conclusion
In this work, self-adaptive firefly algorithm (SAFA) is employed for optimal placement of UPFC in transmission network. SAFA seeks the optimal positions and parameters of UPFC for their installation in IEEE 14, 30, and 57 bus systems. e simulations are performed on MATLAB software for three single objectives (cases) of real power loss minimization, voltage profile improvement, and voltage stability enhancement. e solutions are presented in terms of P loss , TVD, MVSI, and lower and higher voltage magnitudes for three cases of three test systems with comparative results in Section 4. SAFA offers less computational effort, avoiding the suboptimal solution and convergence enhancement. It is evidenced from the solutions that SAFA renders healthier performance than HBA and BFA for UPFCs placement with appropriate parameters; thereby, power system can be operated in reliable and secured manner. e proposed SAFA approach owing to its simple computations has been able to provide practical implementation of any size of power system. is study opens several lines for future work. Analysis of the existing tradeoffs and evaluation of other metrics are some of the future works of this research. is work can be extended by following the same fashion for large-scale bus systems. Other performance parameters can also be analysed in future works. Various improvements of the proposed SAFA for engineering applications can also be investigated and compared in detail in the future.

BSVC:
Susceptance of SVC g ij : j conductance of line between buses i and j g(x, y): Equality constraints h(x, y): Inequality constraints LI r : Light intensity of the rth firefly nload: Number of load buses ngen: Number of generator buses nd: Number of decision variables P loss : Net transmission loss P sp : Specified real power at PV and PQ buses P(V, δ): Set of real power expressions at PV and PQ buses Q min Gi , Q max Gi : Minimum and maximum reactive power generation by the ith generator, respectively Q Gi : Reactive power generation at ith generator