Signal processing techniques for on-line partial discharge detection and classification

Abstract

Partial discharge (PD) detection plays a fundamental role in monitoring the health of medium voltage (MV) systems. This paper presents a method for PD detection and source recognition in MV sub-stations based on a combination of signal processing techniques. Firstly, PD detection and signal conditioning is carried out. Then, PDs of different sources are separated and finally classified by means of the extension set theory. The obtained results show a classification effectiveness of 100% on single source PDs and an effectiveness of 72.5% in multisource PDs, where PDs from many sources are captured in the same data set.