GrC and ANN based fault diagnosis method of distribution network

Abstract

In order to improve the accuracy and efficiency of fault diagnosis in distribution network, this paper puts forward an intelligent hybrid diagnosis system, which combines granular computing theory with neural network theory, to make the best use of rules reduction of granular computing and fault-tolerance learning capabilities of neutral network. In this paper, the concepts of relative granularity and significance of attributes based on binary granular computing are proposed to select reasonable input variables to form the most simplified rules, and are used as the heuristic information of reduction algorithm. The most simplified rule sets are called to make modeling and parameter identification with BP neural network, and then learning training is done by training samples. Compared with the result of fault diagnosis for one distribution network, it shows that the method can reduce the learning training time, improve accuracy of diagnosis and have better fault-tolerance.