Otimização multi-objetivo de redes de distribuição utilizando algoritmos genéticos visando melhoria da confiabilidade

Abstract

This work proposes a way of modeling electricity distribution networks for reliability analysis, and the optimization of these networks regarding their reliability indices and associated costs using genetic algorithms. The methodology involves considering the distribution networks as graphs, being modeled by convenient matrixes that are useful for predicting how the protection and switching schemes installed in these networks will operate in the case of faults. This way, reliability indices can be calculated, allowing the evaluation of these indices compared to the total cost of the network, which covers installation and maintenance of equipment and de-energization costs associated with the network’s customers. The topology may then be optimized using a genetic algorithm, which permits the improvement of reliability indices in relation to the costs involved.