Efficient digital signal processing techniques for induction machines fault diagnosis

This paper investigates recent advances on modern digital signal processing techniques for induction machines fault diagnosis. An intensive research has been performed in order to improve performances of fault diagnosis techniques by applying enhanced signal processing methods during past few years. Since non-invasive sensors offer relatively simple and cost effective fault diagnosis capabilities, more emphasis is given to stator current analysis rather than vibration or acoustic analysis for electrical machines. Here, further interests have been paid on modern signal processing techniques with a special attention to their performances in time domain, frequency domain and time-frequency domain. A comprehensive review is done on recently developed methods which are applied to the stator current collected from induction machine based test-rigs with electrical and/or mechanical faults. It will be demonstrated that numerous techniques have been adapted to induction machines diagnosis. They have been developed primarily based upon basic digital signal processing techniques in order to achieve a more reliable identification and quantification of fault indexes.