Statistical Analysis of Partial Discharges from Electrical Trees Grown in a Flexible Epoxy Resin

Abstract

Electrical treeing is a long-term degradation mechanism in polymeric insulation, which can lead to electrical failure of HV insulation systems. The rate at which trees grow across the insulation depends on the PD activity occurring within them and hence the detection of the onset of electrical treeing could be established by PD monitoring. In this paper, a statistical analysis of the partial discharges detected during the growth of trees in an epoxy resin will be reported. The aim of this work was to provide additional insight into the physical mechanisms that lead to the observed fluctuations in the partial discharge activity. The results demonstrate interesting correlations between a number of statistical parameters, such as average discharge magnitude and standard deviation in the partial discharge amplitudes. These correlations could also be related to physical parameters such as the applied voltage magnitude and the measured power dissipation due to the partial discharges occurring during tree growth. The implications of this work for deterministic methodologies for the simulation of tree growth as well as for condition monitoring using feature recognition strategies for the early detection of tree growth will be discussed.