Classification of external and internal PD signals generated in molded transformer by neural networks

It is difficult to classify external and internal partial discharges in molded power transformer. To solve the problem, a new classification method by NN proposed. In order to simulate partial discharge source, as internal PD, solid insulator with void is used. And gap air discharges with needle-plane electrode is adopted as external PD. From the experiments, statistical parameters are derived from Φ-q-n pattern. And then, the parameters are used for classification by neural network. It is shown that this method can be useful tool to classify the internal and external PD.