Off-line and On-line Automatic and Unsupervised Partial Discharge Diagnostics of Electrical Asset Components

Abstract

This paper deals with applications of an innovative approach to on-line and off-line partial discharge (PD)-based condition maintenance, where PD detection and analytics are fully automatic and do not require the support of PD experts. PD was measured off-line on a 50 kV A, 22/0.4 kV distribution transformer and on 115 and 230 kV transformer bushings, showing in both cases that PD can be detected and analyzed automatically according to the SID (Separation, Recognition, Identification) procedure. On-line measurements performed on MV generators are also reported and discussed, highlighting that there is potential for on-line MV (and HV) monitoring with automatic and unsupervised noise rejection, and identification of PD phenomena that can feed health assessment software to reach optimized maintenance plans.