Classification of PD Signals Generated in Solid Dielectrics by Neural Networks

모의결함을 갖는 고체절연재에서 발생하는 부분방전 및 패턴분류

The recognition of PD(Partial Discharge) phenomenon is useful for classification of defects. The distribution of stochastic parameters which consisted of those PD pulses data and pulses train can show discriminable characteristics of PD sources. But it is not sufficient to discriminate among to PD sources. In this paper, we suggests that classification method of PD source by NN(Neural Networks) are good tools for differentiate of those. The learning scheme of NN is (Back Propagation learning algorithm(BP).