Statistical analysis techniques for Partial Discharges measurement under DC voltage

Abstract

Partial discharges (PD) phenomenon for HVAC (High-Voltage-Alternating-Current) systems has already been widely studied in literature and from this it has been possible to produce a regulatory document that provides indications for description and identification of the different types of discharges. However, the growing diffusion of HVDC (High-Voltage-Direct-Current) systems makes necessary the development of analysis techniques also for DC case. In this paper data obtained from PD measurement under DC voltage were analyzed through the time-frequency map combined with a density-based clustering algorithm. The results show that, with this approach, it's possible to perform a noise rejection and recognize PD pulses of different sources.