Intermittent fault identification for permanent magnet AC drives based on the Short-Time Fourier Transform

Abstract

Prognosis for failure of an electric machine can be achieved through the detection of noncatastrophic faults. As the frequency of these types of faults increases, the working life of the machine is decreased, leading to eventual failure. In this work, two types of stator faults are studied. The methods developed are based on analysis of the short-time Fourier transform of the field oriented machine currents. Linear discriminant analysis is used to classify between the fault types.