Reconfiguração de redes de distribuição utilizando algoritmos genéticos considerando variações de carga

Abstract

In electrical distribution networks the load is the main element that needs more attention as it changes always throughout the day. The main objective of this work presents a methodology using the optimization technique of genetic algorithms to reconfigure the distribution network in normal state operation, changing the load throughout the day. The best configurations are desired, which reduces the power loss in conductors of distribution networks. So, to attend the load variations and prevent frequently changes in system reconfiguration, a load steps analysis is used, which was analyzed only the hour that represents higher consumption at every level. Thus, typical periods of consumption variation were considered, where the values in this periods are assumed as the maximum value of consumption present in this load scenario. The optimization that find the reduced power loss was addressed using genetic algorithms, exploring the main resources of this method such as selection, recombination genetic operators and mutation. Case studies are presented to demonstrate the methodology and the performance of genetic algorithm, where comparisons between the system with a fixed configuration, without modification in the topology of feeders, and with reconfiguration considering the load variations according with the involved consumption class are realized.