Romero Lázaro, Rubén Augusto [UNESP]Souza, Simone Silva Frutuoso de [UNESP]2017-03-302017-03-302017-02-24http://hdl.handle.net/11449/150038Neste trabalho são apresentados três algoritmos imunológicos artificiais para resolver o problema de reconfiguração de sistemas de distribuição de energia elétrica considerando demanda fixa e também demandas variáveis. Para resolver este problema foram utilizados os algoritmos CLONALG (Clonal Selection Algorithm), Copt-aiNet (Artificial Immune Network for Combinatorial Optimization) e Opt-aiNet (Artificial Immune Network for Optimization). O problema de reconfiguração considerando demandas variáveis é um problema complexo e de natureza combinatória, que tem por objetivo identificar a melhor topologia radial para um sistema de distribuição de energia elétrica, de modo a minimizar os custos das perdas de energia ao longo de um período de operação. Para avaliar a factibilidade em relação às restrições de operação dos sistemas de energia elétrica, e calcular as perdas ativas para cada demanda foi utilizado o algoritmo de fluxo de carga radial de varredura. Os algoritmos foram escritos na linguagem C++. Para avaliar os algoritmos propostos foram realizados testes com os sistemas elétricos de 33, 84, 136, 417 e 10477 barras. Os resultados foram comparados com os existentes na literatura, de modo a verificar a eficiência e robustez dos algoritmos.This work presents three artificial immune algorithms to solve the reconfiguration problem of electrical distribution systems (EDSs) with fixed demand and variable demands. To solve this problem were used the algorithms CLONALG (Clonal Selection Algorithm), Copt-aiNet (Artificial Immune Network for Combinatorial Optimization) and Opt-aiNet (Artificial Immune Network for Optimization). The reconfiguration problem with variable demands is a complex problem having combinatorial nature and aiming at identifying the best radial topology for an EDS, in order to minimize the cost of the energy losses in a given operation period. To evaluate the feasibility of the topology regarding the operating constraints of the EDS and to calculate the active power losses for each demand level it was used a specialized sweep load flow for radial systems. The mentioned algorithms were implemented in C++. To evaluate the proposed algorithms were performed testing using test systems with 33, 84, 136, 417 and 10477 nodes. The results were compared with the ones in the literature in order to validate the efficiency and robustness of the algorithms.porReconfiguração de sistemas de distribuiçãoDemandas variáveisSistemas imunológicos artificiaisCLONALGCopt-aiNetOpt-aiNetReconfiguração de sistemas de distribuição de energia elétrica considerando demandas variáveis utilizando algoritmos imunológicos artificiaisDistribution systems reconfiguration considering variable demands using artificial immune algorithmsTese de doutoradoAcesso aberto00088325533004099080P07303300747184265