Partial discharge distribution pattern analysis using combined statistical parameters

Partial discharges (PDs) in high-voltage insulating systems may originate from various defects. It is important to identify the type of these defects, e.g. internal discharges. surface discharges, corona, etc. In the past decade, PD identification research work has been carried out in a way of extracting fingerprints mainly from phase-resolved and/or pulse-height distribution patterns. In this approach, the important parameters to characterize partial discharges are phase angle phi, PD magnitude q and PD frequency n. However, the complexity of these patterns makes it difficult to find effective parameters for the diagnosis. In this paper, the use of combined statistical parameters derived from various PD distribution patterns are presented and discussed.