Otimização da parametrização de relés em sistemas radiais de distribuição de energia elétrica com utilização do algoritmo genético

• Ano de defesa: 2018
• Autor(a) principal: Silva, Marcos Antonio de Araújo
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Brasil; UFRN; PROGRAMA DE PÓS-GRADUAÇÃO EM ENERGIA ELÉTRICA
• Palavras-chave em Português: Proteção de sistemas; Curtos circuitos; Indicadores de continuidade; Redes de distribuição; Unidades de sobrecorrentes; Filosofia da proteção; CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA
• Link de acesso: https://repositorio.ufrn.br/jspui/handle/123456789/25575

Energy concessionaires are constantly seeking to expand and modernize power systems in face of new consumer demands and requirements established by the National Electric Energy Agency. The protection of electrical systems plays a fundamental role in the maintenance of this system and in the control of its indicators. Distribution systems have several physical and technical characteristics, such as the operating time of the devices, load flows and short-circuit levels. In view of the great diversity of these systems, the construction of a computational algorithm that assists in the mathematical procedures of system parameters, or even that, generate a previous automatic configuration of the devices is necessary in view of the wide demand offered to the protection engineers. The algorithm was initially modeled in order to determine the short-circuit currents in each primary protection device, as well as to calculate the minimum currents of operation of the time-overcurrent units and the timing adjustments of the relays. However, due to the great universe of possibilities of adjustments offered by the code, it became necessary to use the techniques of evolutionary biology in order to choose in this universe of possibilities the individual that makes the parameterization of this system optimized. If in the first generation the chosen individual does not match the optimal function, this population will be evolved from the procedures of natural selection, crossing and mutation of the individuals, present in the evolutionary techniques used in the Genetic Algorithms. It is expected that the genetic coding when applied to microprocessor devices with appropriate time interval characteristics according to standards, the algorithm is efficient and stable. However, for feeder circuits where equipment timing is inadequate for a reliable study of coordination and selectivity, changes in some initial timing parameters will be suggested. If the absolute operation time of the feeder is minimized the procedure performed by the algorithm will be considered valid and the configuration presented at the end will be the best among all of the population.