Feature Extraction Based on the Multiple Harmonic Analysis of 2-Dimensional Pattern of Partial Discharge

Partial discharge will cause deterioration of insulation of high-voltage electrical equipment, resulting in insulation breakdown and short circuit accident. In order to ensure the normal operation of the high-voltage equipment, it is necessary to detect partial discharge online. By using Ultra High Frequency (UHF) defection method, statistical 2-dimensional (2D) patterns of partial discharge are obtained. Extracting the envelopes of the 2D patterns, we do the harmonic analysis for patterns' envelopes, and the harmonic components are used as a characteristic parameter for defect type's recognition. The results show that the harmonic components can be used as characteristic parameter for defect types' recognition. This method has the characteristics of strong real-time and simple operation.