A study of surface partial discharge patterns relevant to the endurance of polymer films by means of a neural network

In this paper the use of a back propagation (BP) neural network, and the problems related to the number and to the type of the data most suitable for its training and validation are discussed. Two differently layered structures are presented, which enable one to realize the classification between different materials and to obtain the correlation of the partial discharge pattern to the relevant degradation time, while examples of the neural network outputs are illustrated.