Description: Induction motors, as critical electromechanical energy conversion devices, demand healthy and reliable operation to ensure production continuity and enhance efficiency. Among the various fault types, bearing faults are the most prevalent, accounting for approximately 40% of all motor failures. Traditional diagnostic approaches rely on vibration sensors mounted inside the motor, which not only increase capital investment and maintenance overhead but also impair practicality. In contrast, stator current signal analysis offers distinct advantages, including convenient signal acquisition, reduced monitoring costs, and facile integration into non-invasive continuous condition-monitoring systems. This paper focuses on outer-race bearing faults in induction motors and conducts an in-depth investigation into the extraction of weak fault features from stator current signals under strong noise backgrounds, leveraging stator current analysis as the primary diagnostic tool.
