Application of Multi-objective Evolutionary Algorithms in automatic restoration of radial power distribution systems

When a permanent fault occurs in a power distribution system, the network can be reconfigured in order to restore the supply of some loads situated on non-faulty paths. This paper presents an algorithm developed in Python for optimize the automatic reconfiguration and restoration of radial power distribution systems after the occurrence of a permanent fault. It uses the Multi-objective Evolutionary Algorithm technique and the Step Method in order to optimize all objectives of a given problem, thus providing a greater number of possible solutions. The goals set to the multi-objective function are the maximization of restored customers, minimization of Joule losses and the number of switching maneuvers in the network for the restoration, which are subject to operational constraints. The software features a set of non-dominated solutions, providing the operator with the option to choose from several effective configurations. The grid is modeled by using the node-depth representation (NDR), and the operating constraints evaluated by the forward / backward sweep load flow method. The 16-bus IEEE test system and a proposed 41-bus test system are used to analyze the response of the developed application, which presents good performance and can be safely used by radial distribution system operators.