Neuro-fuzzy based fault detection identification and location in a distribution network

This paper presents an investigation into neurofuzzy techniques for the accurate detection, classification and location of an electric power fault in a distribution network. Ten different types of faults were studied with respect to a real network. These include: line-to-ground faults (on each of phases A, B and C); line-to-line faults (on phases A-B, B-C and A-C); line-to-line-to-ground faults (on phases A-B, B-C and A-C) and three phase fault. A Mandami-type fuzzy controller was also applied to fault type determination. The results reveal that the developed models detect, identify and locate fault incidences to a high degree of accuracy.