The Fault Diagnosis of Power Transformer Based on Improved RBF Neural Network

The radial basis function (RBF) neural network is prior to BP neural network in the ability of approach, the ability of classification and the rate of train. A fault diagnosis method of power based on the RBF neural network is discussed in this paper. The example shows that two input vectors of different class may be more near than two input vectors of the same class. In order to overcome this defect, improve the ability of approach and the ability of classification, the input data is processed according to data reliability analysis and the center of RBF is trained according to the class of input data. The effect of improvement of RBF network has been approved in the fault diagnosis of power transformer.