An Intelligent Defects Identification for Motors Using Partial Discharge Signals

Insulation material for power systems subjected to long-tern operation may deteriorate because of various reasons. If equipment are poorly maintained, the power system may collapse and cause blackouts. To protect high-voltage power system equipment, identifying and analyzing the partial discharge phenomenon caused by insulation deterioration has become an important approach for preventive fault diagnosis. Thus, integrating PD measurement with signal analysis facilitates identification of the equipment's insulation status, which prevents the equipment from failing without warning, and improves the stability of power supply quality. In this paper, a defects identifications system of high-voltage motor stator's winding insulation for partial discharge was developed. First, the methods used to measurement PD signal were examined and a high frequency current transformer was employed to measure the PD signals. A High-Voltage Motor Stator's Winding was selected as the experimental model. Subsequently, PD tests were conducted on the target by using the pressurization process proposed by IEC 60034-27. The PD analyzer developed in this study can receive the voltage signals of difference sensors and transfer the voltage signals to the 3D PD pattern. Furthermore, this study adopted fractal theory and statistical energy analysis to extract relevant features from the various spectra. Finally, extension theory was applied in the defects identifications system of high-voltage motor stator's winding insulation.