Classification and localisation of multiple partial discharge sources within a high voltage transformer winding

Abstract

Partial discharge (PD) analysis is widely adopted for assessing the condition of the insulation systems within high voltage (HV) transformers. Different PD sources have different effects on the insulation condition of HV transformers. In a typical field environment, multiple PD sources may occur in HV transformer simultaneously. Therefore, source classification is very important to identify the types of defects causing discharges in a HV transformer. In recent years, several classification techniques have been proposed for application in PD analysis. This paper proposes automatic techniques to classify and localize multiple PD sources within a HV transformer winding. The proposed processing technique relies on the assumption that the PD pulses generated from different defects exhibit unique waveform characteristics. Surface and void discharges which are the common types of defect events that may occur within HV transformer windings have been experimentally generated. Each pair combination was injected simultaneously into different locations along the HV transformer winding with analysis of two wideband radio frequency current transformers (RFCTs) data captured from each end of the winding. After PD pulses extraction and wavelet analysis, this paper presents two approaches using two different methods to accurately locate multiple PD sources within an HV transformer winding. The performances of the two approaches for this type of application are presented.