Development of current noise cancellation method using sample delaying for bearing fault diagnosis in induction motors

The dominant components in the stator current of a typical induction motor contain a substantial amount of information that is not related to bearing faults and can be considered as "noise" to bearing fault detection. Due largely to poor signal to noise ratio, the detection of incipient faults in bearings by current signal is still challenging. To improve this ratio, this paper proposes a current noise cancellation method based on non-Gaussian residues for bearing fault diagnosis in induction motors. Experimental results for inner/outer raceway bearing defects verify the merits and effectiveness of the proposed method. In addition, it is experimentally shown that the proposed method is superior to the classic motor current signature analysis and the recent proposed method based on Teager-Kaiser energy Operator on motor current signal. To verify the existence of outer/inner raceway defects throughout experimental results, envelope analysis of vibration signal is used as well.