Incipient bearing fault detection via wind generator stator current and wavelet filter

Bearing faults constitute a significant portion of all faults in rotating machines, including wind turbine generators (WTGs). Current-based bearing fault detection has significant advantages over traditional vibration-based methods in terms of cost, implementation, and system reliability. This paper proposes a new wavelet filter-based method for incipient bearing fault detection using electric machine stator currents. The proposed method can dramatically increase the signal-to-noise ratio (SNR) of the bearing fault related signals in the stator current samples. The normalized energy of the wavelet-filtered stator current signals is mainly related to bearing faults and is applied as the index for bearing fault detection. Experiments are carried out for an induction machine with developed bearing faults; the results show that the proposed method is effective to detect the bearing faults at an early stage.