A Data-Driven Diagnosis Method for Bearing Fault Using Harmonics of Stator Current

Most of the diagnosis methods for the asynchronous motor bearing faults are based on vibration signals. However, vibration signals require additional sensors and extra space for installation. In this paper, a data-driven method for asynchronous motor bearing fault diagnosis based on stator current is proposed. The harmonics of stator current are extracted as features after processing by wavelet denoising and quasi-synchronous sampling algorithm, and the mapping relationship between features and labels is obtained by random vector functional link. The experimental test results show that the algorithm can accurately distinguish healthy bearings and inner race faulty bearings or outer race faulty bearings under different working conditions.