Fault diagnosis of power transformer based on ellipsoidal basis functional neural network

An ellipsoidal basis function (EBF) can make the partition of input space and make a limitary and bounded. Comparing with the Gaussian function of radial basis function (RBF) neural network, the EBF can make the partition of input space more specific. So, it has the higher capability of pattern recognition. A new method to diagnose power transformer faults based on EBF neural network was proposed. First, the data of five characteristic gases obtained by the gas-in-oil analysis were preprocessed, and 6 characteristic data for fault diagnosis were distilled. Then, the EBF neural network was trained with the sampling data from the above process. Final, the testing data were respectively identified by the EBF neural network and RBF neural network. The test results show that the EBF neural network has the higher rate of fault diagnosis of power transformer than that of RBF neural network.