Integrated diagnosis techniques of high voltage induction machine with rotor winding faults

This paper analyzes the novel techniques adopted to diagnose machine rotor winding fault. In order to accurately extract the fault characteristic in stator-steady current, frequency subdivision is used to enhance the frequency resolution of the signal and self-adapting filtering is used to enhance the magnitude resolution of the signal. Particular frequency component in the starting current is used for fault diagnosis, this greatly simplify the fault diagnosing technique based on starting current and make it applicable. Residual voltage after AC dump of the stator winding is used for fault diagnosis, this can effectively eliminate the effects caused by unsymmetrical source, load fluctuation, and iron saturation. The application of all these techniques can effectively diagnose the incipient faults of high voltage induction machine. Finally this paper introduces the basic structure and the main performance of the realized high voltage machine fault diagnosis system.