Comparative Analysis of partial Discharge Patterns from Different Artificial Defects by Mean of Conventional Phase-Resolved Partial Discharge Analysis and a Novel Chaotic Analysis of Partial Discharge

Abstract

Partial Discharge (PD) phenomena occurring from an insulating defect in a large power apparatus can be regarded as a deterministic dynamical process in which a PD must occur if the local electric field is to be sufficiently high in any given circumstances. In this respect, a new analytical method based on Chaos theory, called `Chaotic Analysis of Partial Discharges', is proposed for the purposes of this work. Throughout this paper, qualitative and quantitative information relating to the PD signals from different insulating defects is analyzed by using a 3-dimensional attractor in a phase space, the fractal dimensions, the Lyapunov spectra, and the Kolmogorov-Sinai entropy. These results are compared with those from conventional phase-resolved partial discharge analysis (PRPDA) involving the same kinds of defects. It can be said that the PD phenomena in an insulating system is not merely a random process but a chaotic process, possessing obscure rules behind its randomness. Furthermore, it should be pointed out that the proposed method should be employed to determine the physical degradation process due to PD in an insulating system.