Separation of multiple partial discharge sources within a high voltage transformer winding using time frequency sparsity roughness mapping

Partial discharge (PD) measurements can evaluate integrity of transformers' condition insulation system. In high voltage (HV) power transformers, the winding system consists of multiple dielectric media all of which can degrade and subsequently exhibit pre-breakdown behavior, the ability to accurately separate between different PD signals generated from different sources is seen as an important function of diagnostic systems. This paper is concerned with the feasibility of locating two types of PD sources; surface and void discharges simultaneously into a HV transformer winding. Based on the fundamental theory using the theory of travelling waves along passive transmission lines, PD produces a signals that will propagate towards both ends of the transformer winding - the bushing and neutral to earth connection point from the source. Thus this paper is based only on measurement data from two wideband radio frequency current transformers (RFCTs) placed at the bushing tap-point to earth and neutral to earth point. After PD pulse extraction and mathematical morphology (MM) decomposition, time frequency sparsity roughness mapping is applied followed by density-based clustering of application with noise (DBSCAN) to separate multiple PD sources and estimate their location.