Based on fuzzy rough sets and GA-BP neural network method of distribution network fault line

According to the BP neural network fault line when the input data amount is large, its structure is complex, convergence slowly, and easy to fall into the local optimal shortcomings, we will put fuzzy rough sets and the genetic algorithm to optimize the method of neural network into one-phase ground fault distribution network in line. We obtained the line of zero sequence current signals through the simulation tests and a variety of extract characteristic information fusion. We use rough sets theory to attribute reduction conditions and remove redundant condition attribute. We will reduction attributes as input layers of the BP neural network. And then we through the genetic algorithm to optimize the BP neural network was trained and tested. The results show that this method has the more training speed and lower false positives than the traditional method. And the system can meet the power requirements for precision and accuracy.