An efficient method for sizing and allocation of distributed generation and voltage regulators in a distribution network Saeedeh Ketabipour, Shahrokh Shojaeian * Department of Engineering, Islamic Azad University, Khomeinishahr Branch, Isfahan 84175-119, Iran Corresponding Author Email: shojaeian@iaukhsh.ac.ir https://doi.org/10.18280/mmc_a.910209 Received: 27 May 2018 Accepted: 30 June 2018 ABSTRACT One of the well-known challenges in designing long distribution lines is to determine the efficient locations of voltage drop compensating devices e.g. Voltage Regulators (VRs) and Distributed Generations (DGs). Determination of optimal locations, results in more economical benefit of these equipments. The majority of the works presented in the literature try to satisfy this goal, using classical load flow methods, such as Newton- Raphson approach or the back-forward method. However, these classical methods may suffer from some serious deficiencies. Namely, the first method is faced to the risk of divergence in distribution systems and the second is time consuming. Moreover, both methods impose quite significant complexity. The method presented in this paper employs a load distribution technique in addition to the genetic algorithm solution, enabling the investigation of simultaneous effects of the distributed generation and voltage regulator on a distribution feeder. In the proposed algorithm, an objective function composed of power losses and voltage deviations is used to obtain an optimal location of the above mentioned equipments (DGs and VRs) and the efficient size of the distributed generation system. The proposed idea has been examined for the IEEE 33-Bus test system and its favorable efficiency is confirmed. Keywords: distribution network, voltage regulator, genetic algorithm, voltage drop, distributed generation 1. INTRODUCTION Power distribution network is an interface medium between power consumers and transmission systems. The two issues of voltage drop and power losses have been known as the main challenges which cause serious problems in distribution systems, especially when the loads are concentrated at the end of long feeders. The low voltage level used in power distribution networks implies the flow of quite high currents. This significantly increases the value of power losses in the distribution networks compared to the transmission networks. Various methods are suggested to reduce the power losses in distribution systems. Some of the most prevalent methods included capacitor siting, efficient placement of DGs and reconfiguration of the network and voltage regulators. Some researches show that application of DGs with unfit capacity or placing them at inappropriate locations within the network may increase the power losses, compared to the state before the installation of DGs [1]. Various optimization methods have been proposed in the literature, which can be classified into analytical methods [2] and heuristic methods [3- 9]. Heuristic methods include the evolutionary optimization approaches based on the Genetic Algorithm (GA) [3], Artificial Bee Colony (ABC) algorithm [4], fuzzy research reasoning approach [5], Particle Swarm Optimization (PSO) [6], Ant Colony algorithm [7], Cerebellar Model Articulation Controller (CMAC) [8], Clonal selection algorithm [9] and Leaping Frog algorithm [10]. A combination of genetic algorithm and PSO is used in [11] to determine the optimal sizes and places of DGs in order to reduce the losses and voltage drop over the network. In [12] the genetic algorithm is employed to determine the optimal sizes and locations of DGs, which not only minimize the power losses and voltage drops, but also enhance the system reliability. The genetic algorithm is also applied in [13] to determine the sizes and locations of DGs which minimizes the costs and losses in the distribution network. In addition to Using the genetic algorithm to simultaneously find the optimal locations of DGs and VRs, this paper also attempts to determine the optimal sizes of DGs, which minimize both power losses and voltage drops in the network. All of [11-12] and [13] employ the genetic algorithm just to find the optimal locations of DGs. However, none of them considers the locations of VRs as variable parameters to be found through the optimization process. However, our proposed method jointly optimizes the sizes and locations of DGs and the locations of VRs, to minimize power losses and voltage drops in the distribution system. Many papers have investigated the use of capacitor banks in order to minimize the power losses and optimize the voltage levels within the network [14-15]. However, there are very limited papers which have also considering the role of VRs in this problem. Therefore, bringing VRs into consideration may provide some noteworthy opportunities for new models and algorithms. In [16], a computerized algorithm is used for controlling the optimized voltage levels, which is suitable for large radial distribution networks. In [17] a two-steps algorithm has been proposed for optimal placement of VRs in distribution systems. In the first step, voltage regulators are placed at candidate buses (and the tap position is determined), aiming at minimizing voltage drops and real power losses. In the second step, an attempt to reduce Modelling, Measurement and Control A Vol. 91, No. 2, June, 2018, pp. 83-88 Journal homepage: http://iieta.org/Journals/MMC/MMC_A 83