Application of statistical analysis in result repeatability evaluation of acoustic emission measurements generated by partial discharges

Abstract

The subject matter of this paper deals with the improvement of the acoustic emission (AE) method in diagnostics of insulation systems of power appliances. The paper presents the results of the statistical analysis of the AE pulses generated by the setups modeling basic partial discharge (PD) forms that can occur in oil insulation systems. Especially, the results of parametric and non-parametric tests of quality of fit are presented that were carried out in order to determine the probability distribution of the AE pulses measured, to test the repeatability of the frequency analysis results obtained, and to check the influence of the factors that can disturb this repeatability. The statistical analyses were carried out based on the values of three descriptors, i.e. shape coefficient, peak coefficient and median frequency, which make the identification of basic PD forms occurring in insulation oil possible.