Hilbert Huang Transform and pattern recognition to detect defects in induction motor

This paper presents recent time-frequency analysis technique called Hilbert Huang Transform (HHT) used for non-stationary and nonlinear signals. This technique is applied in combination with a pattern recognition method to detect defects in induction motor. Studied defects in our case are one broken rotor bar, two broken rotor bars and broken ring on the rotor. The aim of this study is to improve the classification results of these different rotor faults. The classification results obtained by using the HHT method and those obtained by using the temporal signals are compared.