ANN based three stage classification of arc gradient of contaminated porcelain insulators

In this paper, the transition from weak inception current flow on the surface of the contaminated porcelain insulators till flashover occurs, is classified into three stages which can better be explained in terms of arc voltage gradient. The more popular Ayrton´s equation is chosen which computes arc voltage gradient in terms of arc current I and Ayrton´s constants A and n. The present work describes the development of a multi layer Feed Forward Neural Network (FFNN) classifier model using backpropagation algorithm for training, to discriminate the arc gradient for the three stages considered, for the given values of A, n and I as the input parameters. The model is tested and the results show that, Neural Network structure with six nodes in the hidden layer is best suited for the present classification. The percentage of correct classification is found to be 100 in all the three classes.