Design of Pattern Recognition Application of Partial Discharge Signals Using Artificial Neural Networks

Abstract

This paper presents the experimental result of detecting Partial Discharge (PD) for three different sensors and an application for recognizing PD type using Artificial Neural Networks (ANN). Nowadays, Artificial Intelligence is being used by various types of researchers, included in the field of Electrical Power Engineering. Therefore, the authors utilize one of its branch for doing a better diagnosis in High Voltage Equipment which is Artificial Neural Networks. This method was chosen because it has been proven to have a high level of accuracy on pattern recognition. Database compilation is done by taking PD data with PD sources in the form of corona and PD in oil insulation at three different voltage levels. The recognition of PD patterns is done by looking at the PD signal phase pattern while the PD signal assessment is done by looking at the maximum amplitude for both positive and negative PD and also the number of PD appearances in each cycle. This important information will be used in the process of making ANN, then the network will be used for pattern recognition and assessment of PD signals in the application made.