Sistema de Inferência Fuzzy Multicamadas Aplicado à Localização de Faltas em Sistemas de Distribuição usando Apenas Medições de Tensão
Lucas Assis de Moraes1; Rogério Andrade Flauzino1; Marcel Ayres Araujo1; Oureste Elias Batista1
1 Universidade de São Paulo
doi:10.20906/CPS/SBSE2016-0347
Resumo
This paper aims to develop a multilayer fuzzy inference system (FIS) to locate faults in a distribution system. The FIS creates optimized rule base and membership functions by adjusting each parameter using the descending gradient method. The FIS is trained using simulated fault data from the low-side voltage of transformers in a real distribution system. Real and imaginary parts of phase, line and sequence voltages are used. It is verified that after training, a good answer of the location of the fault is only possible to be achieved if there are a great number of measurement points in the grid.
Palavras-chave: Distribution System; Fault Location; Fuzzy Inference Systems