Bearing Fault Diagnosis Under Time-Varying Speed and Load Conditions via Speed Sensorless Algorithm and Angular Resample

Bearing faults are one of the major causes of motor failures. The vibration or current signals are the most commonly used in bearing fault diagnosis (BFD). However, their applications are also limited by some disadvantages. This study addresses a novel BFD method based on speed sensorless observer and angular resample(AR). Based on the analysis of the effect on the rotor speed signal caused by bearing faults in permanent magnetic synchronous machine (PMSM), a speed sensorless observer is designed to obtain the rotor angle and speed for reducing the cost and maintenance of the optical encoder. The angle and speed signal are pre-processed to fit the demand of the AR method, and the angle signal is used to resample the non-stationary speed signal into a stationary one in angular domain for the order spectrum analysis. The existence of excitation at the bearing fault characteristic order in the spectrum of the resampled signal is used for fault diagnosis. Experimental studies are carried out in both speed-varying and torque-varying conditions to demonstrate the effectiveness of the proposed method for the detection of bearing faults of PMSM.