Multi-objective service restoration of distribution systems using group-based two-stage methodology

Abstract

This paper develops a two-stage methodology tailored for multi-objective service restoration in large-scale, unbalanced distribution systems. The developed methodology combines optimization techniques with a ranking-based switch search method for efficient and robust performance. It is designed to explore the merits of meta-heuristics and local heuristics. The group-based Particle Swarm Optimization (PSO) algorithm in Stage I is designed for global search, providing several representative solutions for the following stage, which is designed for refining the solutions of Stage I. The heuristic method in Stage II can quickly improve upon the solution provided by the PSO. Multiple solutions obtained in Stage I are useful for system operators since the best solution suggested by the proposed method needs to be compromised with engineering judgement. Operators can choose a plan (solution), more suitable to real situations from the solutions provided by the two-stage methodology. Evaluation results on IEEE 123 node test feeder show the effectiveness of the proposed methodology.