Otimização dos parâmetros dos controladores de um filtro ativo de potência paralelo utilizando algoritmos multi-objetivo

• Ano de defesa: 2018
• Autor(a) principal: Santana, Nelson Henrique Bertollo
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Espírito Santo; BR; Mestrado em Energia; UFES; Programa de Pós-Graduação em Energia
• Palavras-chave em Português: Redes elétricas; Correntes elétricas; Engenharia/Tecnologia/Gestão; 620.9
• Link de acesso: http://repositorio.ufes.br/handle/10/10603

The increment in utilization of non-linear loads by the grid users is a phenomenon that has begun in the middle of 20th century and has increased year after year. That kind of load has as main feature the harmonic current emissions due to its energy conversion process, which causes many nocive e ects in the grid, like conductors and transformers heating, equipment malfunctioning, including the protection ones, among others. In the range of available alternatives to overcome the harmonics problem, the active power lters are a good choice, due to its good operational performance, specially for low order harmonics. In this context, many control techniques have been developed for these kind of lters operation, as hysteresis control, deadbeat control and the linear current control techniques, like PI-SRF, P-MRI and P-SSI. The last ones have the advantage of operating with well de ned switching frequency and independence of the correct system parameters information to its good performance. However, the tunning of the controllers is a relevant problem to its utilization. Thus, this work proposes the utilization of otimization algorithms for the controllers tunning of two di erent control techniques: PI-SRF and P-SSI. To this end, two di erents metaheuristics, MOPSO and MOABC, were used so it may be possible to compare the performance of each one for tunning each technique. The results shown the superiority of MOPSO, achieving 17% better results and spending 50% less computacional time when compared to MOABC. The best parameters were used in a gain scheduling simulation, wich reduced the source THD to 1.00% with a 20.19% THD connected load.