Simultaneous optimization of phase balancing and reconfiguration in distribution networks using BF–NM algorithm R. Hooshmand a,⇑ , S.H. Soltani b a Electrical Engineering Department, University of Isfahan, Isfahan, Iran b Electrical Engineering Department, Islamic Azad University, Najaf Abad Branch, Isfahan, Iran article info Article history: Received 4 August 2011 Received in revised form 8 February 2012 Accepted 12 March 2012 Available online 22 April 2012 Keywords: Phase balancing Rephasing strategy Reconfiguration technique BF–NM algorithm Distribution networks abstract Rephasing strategy is one of the main methods used for phase balancing and neutral current reduction in electrical distribution networks and the reconfiguration technique is an effective method for network loss reduction. In this paper, a new method for the simultaneous implementation of reconfiguration and phase balancing strategies is presented as a combinational strategy. In order to solve the proposed opti- mization problem, Nelder Mead algorithm combined with a bacterial foraging algorithm (BF–NM) is used based on a fuzzy multi-objective function. The proposed method allows for the simultaneous execution of reconfiguration and phase balancing while minimizing the interruption cost of rephasing in addition to eliminating network unbalancing and reducing neutral current and network losses. To demonstrate the efficiency of the BF–NM algorithm, its performance is compared with bacterial foraging (BF), particle swarm optimization (PSO), genetic and immune algorithms (GA and IA). The proposed method is applied to the IEEE 123-bus test network for evaluation. The simulation results confirm the efficiency of the method in reducing the system costs and network phase balancing. Crown Copyright Ó 2012 Published by Elsevier Ltd. All rights reserved. 1. Introduction One of the most important problems in electrical distribution networks is current unbalancing in feeders. Various problems are caused by this unbalancing; chiefly loss increasing, phase voltage unbalancing, over-current relay action and over-loads in distribu- tion network equipment. Rephasing strategy is one of the best solutions to this problem, and different methods are available for implementing it. Rephasing strategy was first introduced in 1997 as a mixed integer programming technique [1]. The high computational time is one of the problems of this technique, which has limited its use to small feeders. So, researchers have attempted to solve this problem using intelligent methods, e.g., simulated annealing (SA) but have not yet solved the problem of its high computational time [2]. Therefore, faster intelligent methods must be used for phase balancing [3–6]. In Ref. [3], phase balancing is accomplished using GA, which can reduce computation time and solve phase unbalanc- ing problem to a large extent. In this method, the number of reph- asing operations is high, which increases the rephasing cost. In Ref. [4], rephasing strategy is executed by PSO algorithm. In Refs. [5,6], rephasing is performed by expert system and immune algorithm. The rephasing cost minimizing problem is solved by including interruption and labor costs in the objective function. Additionally, Ref. [7] introduces software for phase swapping using simple fee- der selection based on a graphical user interface application. Rephasing strategy performs well in phase balancing but its capability for loss reduction is low. Reconfiguration technique is one of the main methods used to minimize losses in distribution networks. In reconfiguration technique, by choosing the appropri- ate state corresponding to the changing status of available switches in a network, the network configuration is altered so as to minimize network losses by retaining the radial structure of the network. In this regard, different strategies are presented. Included among the conventional methods are constructive [8] and destructive [9] methods. In the constructive method, it is assumed that all of the switches are first open; they are then closed one by one, and the net- work limits are checked. In the destructive method, it is assumed that all of the switches are first closed, and the optimum network structure is obtained by opening a specified number of switches. As expected, intelligent methods show improved capabilities over conventional methods. Among them, it can be pointed to the follow- ing: Ref. [10] is the combined use of ant colony and immune algo- rithms. In [11], GA has been used to solve the reconfiguration problem, where the matroid theory is applied to optimize GA oper- ators. Ref. [12] applies a heuristic method to solve the reconfigura- tion problem. Also in [13], tabu search is used for the problem solution. Ref. [14] presents a comparative study for four modern heuristic algorithms: reactive tabu search, tabu search, parallel SA, and GA to service restoration in distribution systems. In Ref. [15], 0142-0615/$ - see front matter Crown Copyright Ó 2012 Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijepes.2012.03.010 ⇑ Corresponding author. E-mail addresses: Hooshmand_r@eng.ui.ac.ir (R. Hooshmand), shirinsolta ni2008@yahoo.com (S.H. Soltani). Electrical Power and Energy Systems 41 (2012) 76–86 Contents lists available at SciVerse ScienceDirect Electrical Power and Energy Systems journal homepage: www.elsevier.com/locate/ijepes