Examination and Determination of Partial Discharge Source using Gaussian Naive Bayes (GNB) Technique

Abstract

The outline of partial discharge (PD) is an substantial instrument for high-voltage insulation systems diagnostics. In different PD data representations, human experts can detect possible isolation defects. PD (PRPD) patterns are one of the most commonly used representations. To ensure the confident operation of HV-equipment, the statistical properties of PDs must be linked to the defect properties and the type of defect determined. Classifier Naive Bayes are a family of simple 'probabilistic classifiers,' which use Bayes theorem with robust presumptions of independence among features. At the moment the GNB's importance in assessing its appropriateness for PD actions is being considered. The model is built in Google's Python co-laboratory and can be generated from the PD source on a graphical user interface, whether it is void, surface or crown discharge, using statistically based parameters. The model is also available for this method.