Based on Rough Set and RBF Neural Network Power Grid Fault Diagnosis

Abstract

In order to improve the accuracy of troubleshooting results and save diagnostic time, rough set and RBF neural network are used to diagnosed fault of power grid.Rough sample reduction program is used to reduce the fault samples, and the reduction decision table is obtained as the input of the RBF neural network, get the training results.The fault samples without rough set reduction are input into RBF for training, and the two training results are compared. It is found that the samples using the rough set reduction have the same diagnostic ability as the initial decision table.Obviously, the rough decision set initial decision table can greatly reduce the size of the training sample and save diagnosis time.