Digital Signal Processing for Induction Machines Diagnosis - A Review

The aim of this paper is to present a wide view of the digital signal processing technique (DSPT) which can help to improve induction machine fault diagnosis. The classical as well as the modern signal processing methods are presented with attention to the performances in frequency, time and time- frequency domain analysis. These methods are applied to different signals coming from the three-phase induction machine sensors under different operating conditions. It is shown that the best diagnostic performances can be obtained when the DSPTs are well adapted to the observed phenomena.assuming that the signal is a set of harmonically related sinusoids.