Distribution network reconfiguration (DNR) continues to be a good option to reduce technical losses in a distribution power grid. However, this non-linear combinatorial problem is not easy to assess by exact methods when solving for large distribution networks, which requires large computational times. For solving this type of problem, some researchers prefer to use metaheuristic techniques due to convergence speed, near-optimal solutions, and simple programming. Some literature reviews specialize in topics concerning the optimization of power network reconfiguration and try to cover most techniques. Nevertheless, this does not allow detailing properly the use of each technique, which is important to identify the trend. The contributions of this paper are three-fold. First, it presents the objective functions and constraints used in DNR with the most used metaheuristics. Second, it reviews the most important techniques such as particle swarm optimization (PSO), genetic algorithm (GA), simulated annealing (SA), ant colony optimization (ACO), immune algorithms (IA), and tabu search (TS). Finally, this paper presents the trend of each technique from 2011 to 2016. This paper will be useful for researchers interested in knowing the advances of recent approaches in these metaheuristics applied to DNR in order to continue developing new best algorithms and improving solutions for the topic.

/pdf/recent-trends-of-the-most-used-metaheuristic-techniques-for-nbrwd8psc0.pdf

Recent Trends of the Most Used Metaheuristic Techniques for Distribution Network Reconfiguration

The purpose of optimal reconfiguration of a distribution network is a matter of finding the best topology of the network to achieve the considered goals. This paper presents an algorithm for optimal reconfiguration of distribution networks using Binary Genetic Algorithm (BGA) based on Graph Theory (GT) as a multi-objective optimization problem. The objective function considers the real power losses, and the network voltage deviation index. Also, lines and transformers loading limits are considered as constraints in the optimization problem. To assess the capabilities of the proposed approach, two networks are studied: the IEEE 33-bus distribution network as a small grid and Bardsir regional medium voltage distribution network as a large-scale grid. Simulation results show both voltage profile improvement and loss reduction compared to the situation before the reconfiguration. The algorithm is implemented in MATLAB software and then evaluated by DIgSILENT Power Facrory 14.0.515 software.

Bardsir network reconfiguration using Graph Theory-based Binary Genetic Algorithm to reduce loss and improve voltage profile

Considering that the information uploaded by fault indicator devices is often lost or wrong in actual operation, a fault diagnosis method for distribution networks based on multi-source information is developed in this paper. The multi-source information includes information from distribution automation terminals and the customer electric information acquisition system (CEIAS). Firstly, using the information uploaded from the remote terminal unit of distribution automation system, the fault hypothesis set is formed based on the immune algorithm. Then the fault hypothesis set is further judged by the information collected from CEIAS to get the correct fault section and the incorrect alarms. The correct conclusion can also be obtained by using this method in case of that the key information is incorrect or there are many incorrect or missing alarms at the same time. The validity and accuracy of the proposed method are verified by the simulation of F1 feeder in IEEE RBTS BUS-2 distribution system.

Fault diagnosis method for power distribution systems based on multi-source information

In this paper, a bio-immune mechanism based genetic algorithm (GA) optimization approach for the reconfiguration of distribution system is proposed. Some mathematical methods such as memorial property, diversity and self-adaptive feature of the natural immune system are introduced which prevent premature problem of GA effectively and acquire global optimum with high speed during the iterative procedure. Moreover, a coding strategy, representing distribution system features, is adopted which codes the switches of the tie switch-decided loop in the same chromosome. The improved coding strategy can effectively prohibit invalid solution, reduce search space and accelerate speed. A 33-bus distribution system is tested and the corresponding results demonstrate the effectiveness and feasibility of the proposed method for application.

Yuzhong Qin

Papers

An immune genetic algorithm based approach for distribution system reconfiguration