[129] Using minimum entropy decomposition and spectral kurtosis for rolling element bearing fault analysis in a permanent magnet motor

The rolling element bearing is one of the most vulnerable components in a permanent magnet motor because it is most often under high-load and high-speed running conditions. Prompt diagnosis of rolling element bearing faults is critical not only for the safe operation of machines, but also for the reduction of maintenance cost. This work studies the diagnostics of bearing inner race and outer race faults. Rolling element bearing fault diagnosis using vibration signals relies profoundly on using the well-developed methodology of envelope analysis. Envelope analysis is traditionally performed through amplitude demodulation of an excited frequency band using Hilbert transformation to determine the fault frequencies. The enveloping technique suffers due to a low signal-to-noise ratio and not knowing the frequency band in which it should be applied. In this paper the authors present the fault analysis results of a bearing with inner and outer race defects in a permanent magnet motor using traditional enveloping combined with advanced signal processing techniques. Minimum entropy decomposition (MED) is used to enhance bearing signals in cases with a low signal-to-noise ratio. Spectral kurtosis (SK) is used to identify the demodulation band. The results show that the efficiency of envelope analysis in providing a correct diagnosis is increased when it is augmented with bearing signal enhancement and frequency band determination techniques.