Motor Bearing Fault Diagnosis Based on Frequency Conversion Estimation and Order Analysis

Abstract Order analysis (OA) has been proven to be an effective tool for motor bearing fault diagnosis under variable-speed conditions, but it is a challenge to perform OA when a tachometer is unavailable to obtain the rotating speed. In this study, a new angular resampling method based on instantaneous frequency estimation of the transient current signal is proposed for bearing fault diagnosis with speed fluctuation. First, the fundamental rotating frequency curve is extracted from the acquired phase current signal of a brushless direct current motor (BLDCM). The mechanical rotating angle curve is then calculated by integrating the fundamental frequency curve. The synchronous sampled vibration signal of the fault bearing is resampled in the angular domain according to the obtained rotating phase information. The effectiveness and efficiency of the proposed method are verified through a BLDCM experimental setup.