Internal Fault Identification and Classification of Transformer with the Aid of Radial Basis Neural Network (RBNN)

This paper deals with the identification and classification of internal fault current of power transformer occurring during the time of abnormal condition. The need of internal fault current classification is to avoid the complexity of the fault category. In this paper, the inrush current and short circuit current of the transformer internal windings are classified from the nominal current. Before the classification process, the analytical model parameters based identification of inrush current is described. The analytical model parameters considered are wave shape and wave peak of the current. The output of the power transformer is applied to classifier and then, the shape and peak of the waveform are extracted from the classifier. Here, an artificial intelligence based radial basis neural network (RBNN) classifier is used to extract the wave parameters. In the RBNN, the Gaussian function is considered as an activation function. The proposed internal fault identification and classification technique is implemented and tested with different ratings of transformer, and the fault classification performances are evaluated. Then, the evaluated results are compared with the feed-forward network.