Time-Frequency Classification Applied to Induction Machine Faults Monitoring

This paper presents a new method of designing an optimized time-frequency representation (TFR) from a time-frequency ambiguity plane applied to the induction machine faults classification. This method is composed of two sequential processes: a feature extraction and a classification. In the process features extraction, the time-frequency representation (TFR) have been designed for maximizing the separability between classes representing the different faults; bearing fault, stator fault and rotor fault. The classification of a new signal is based on the Mahalanobis distance. The diagnosis is independent from the level of load. This method is validated on an 5.5 kW asynchronous motor test bench