1318 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 28, NO. 2, MAY 2013 Two Circular-Updating Hybrid Heuristic Methods for Minimum-Loss Reconfiguration of Electrical Distribution Network Abdullah Asuhaimi Mohd Zin, Senior Member, IEEE, Ali K. Ferdavani, Student Member, IEEE, Azhar Bin Khairuddin, Member, IEEE, and Marjan M. Naeini, Student Member, IEEE Abstract—This paper presents two hybrid heuristic methods for reconfiguration of the radial electrical distribution system. The first stage of the proposed techniques is a circular min- imum-branch-current updating mechanism to bypass the local optimum points. Then, by a circular neighbor-chain updating technique, the best known configuration is obtained. The proposed approaches have been tested on the six well-known electrical dis- tribution networks. The results show that the proposed methods have remarkably achieved the best-known optimum solution with more reliability, accuracy and efficiency than many other published methods. Index Terms—Algorithms, circular updating, heuristic method, losses, minimum current, optimization methods, power distribu- tion, radial system, reconfiguration, switches. NOMENCLATURE Power loss. Set of branches included closed switches. Number of loops. I. INTRODUCTION H AVING optimized the electrical power distribution system, the utility can reduce the total costs of energy to customers. With this respect, reconfiguration of radial electrical distribution network (ROREDN) is one of the outstandingly issues with the aim of minimizing the power system loss as well as the voltage drop, increasing reliability, balancing loads and energy restoration [1]. Minimizing the total resistive power loss is certainly the main objective of ROREDN. Generally, power loss depends on branches’ currents, which are being changed due to variation of loads. Having changed on the status of even one switch, i.e., open to close one or vice versa, some branches’ currents Manuscript received February 02, 2012; revised May 18, 2012 and July 27, 2012; accepted August 28, 2012. Date of publication December 12, 2012; date of current version April 18, 2013. This work was supported by Universiti Teknologi Malaysia (UTM) and Ministry of Higher Education (MOHE), Malaysia under FRGS vote No. 78560. Paper no. TPWRS-00097-2012. The authors are with the Faculty of Electrical Engineering, Univer- siti Teknologi Malaysia, Johor Bahru, 81310 Johor, Malaysia (e-mail: abdullah@fke.utm.my; ferdavani@ieee.org; azhar@fke.utm.my; m.mortazavi. n@gmail.com). Digital Object Identifier 10.1109/TPWRS.2012.2218290 are altered, and therefore, the power loss is changed. In order to calculate the power system loss, many formulas have been improved such as presented in [2]–[4]. As a summary, the problem of ROREDN is basically applied to find the best status of switches to minimize the total power system loss [5]–[12]. There are certain constraints in solving the ROREDN problem. The electrical system at the end of reconfiguration should be completely radial one. Besides, both Kirchhoff’s current and voltage laws must be observed to solve the AC load flow equations of the radial electrical system. Furthermore, the buses’ voltage magnitudes and the branches’ currents should be limited into their boundaries. By considering various com- binations of these limitations, different final solutions may be obtained by the optimization method. There are many approaches, which are categorized as math- ematical, Meta heuristic and heuristic methods, in ROREDN, which is a nonlinear optimization problem. Generally, mathe- matical methods should be followed in special situations such as continuous and deviational limitations. The population and random-base techniques, classified as Meta heuristic methods, usually require large computational times to perform load flow analyses. On the other hand, the heuristic methods (HMs) are usually used in a wide range of the power system simulations, designations and operations because of their appropriate con- vergence speeds in obtaining the optimum solution. So far, many heuristic methods are presented by researchers. Civanlar et al. [3], Baran and Wu [2], and Shirmohammadi and Hong [13] published the results of their groundbreaking studies in the field of ROREDN. Basically, HMs are classified in four groups [1]: Close-all-switches [13], [14], interchange-switches [2], [3], [15]–[19], Open-all-switches [20] and hybrid [21]–[26] strategies. The theory used in [13] is to select the switch carrying the lowest current after performing power flow for opening while other unselected switches are considered as being closed. This method is improved further by Raju and Bijwe [23], by utilizing the objective function (OF) in second stage of their algorithm, and Mohd Zin et al. [14], which is based solely on the current indicator approach as that of [13]. The value of the OF during the progress is not always improved as shown in [14]. As a re- sult, these techniques could not guarantee the global optimum solution. Ferdavani et al. [19] presented the circular neighbor-chain- updating method that any open switch is replaced with its neighbor in condition of improving the OF. However, this 0885-8950/$31.00 © 2012 IEEE