Diagnostic of HVDC systems using partial discharges

Abstract

This paper presents a methodology for the diagnosis of HVDC systems based on partial discharge analysis. In particular, the characteristics of partial discharge phenomena from different test objects are highlighted, with the aim of providing tools enabling automatic identification of the type of PD, as well as noise and disturbance rejection, for on-line and/or off-line PD testing of HVDC equipments. The results show that, using a fuzzy identification system based on higher-order moments of the PD magnitude and inter-time distributions, corona phenomena can be identified easily, whereas it is sometimes difficult to distinguish between internal and surface partial discharge sources. For this purpose, the characteristics of internal and surface partial discharge phenomena as a function of influence variables as humidity and temperature are reported. By exploiting these characteristics through on-line/off-line PD monitoring systems, accurate identification of the PD phenomena affecting an electrical apparatus can be attained.