Automation of High-Voltage Electrical Equipment Insulators Diagnostics by Partial Discharges Method

Abstract

Currently, the method of partial discharges is actively developed to diagnose the insulation of high-voltage electrical equipment of stations and substations. Recently, artificial neural networks have been used to recognize partial discharges and analyze their characteristics and predict the development of insulation defects. To train neural networks, in turn, model partial discharge sources with well-known characteristics are often used. The presented work is devoted to the development of criteria for automatic recognition of the localization of a defect and a pre-breakdown situation in the insulation of high-voltage electrical equipment, based on the analysis of the characteristics of partial discharges.