Neural network application in PD pattern analysis

摘要

The Computerised Discharge Analyser (CDA3) is a PC-based data acquisition system. It captures in real-time individual PD pulses from the output of a conventional analogue discharge detector. Integrated analysis of the raw data is carried out by software. It was shown that automatic recognition of the PD patterns is possible with the use of a neural network. The input to the network can be the bivariate phase-resolved distributions. Its size can be reduced by feeding the network with pre-processed data such as the statistical moments of the univariate distributions or the IEC-270 integrated quantities.