Vibration Signature Analysis for Broken Rotor Bar Diagnosis in Induction Motors Based on Cyclic Modulation Spectrum

Broken rotor bar (BRB) is a common fault in induction motors (IM), which results in devastating consequences in the stator and reduces system reliability if not detected in a timely manner. An effective monitoring method of BRB is desired to guarantee production efficiency and service life of IM. Currently, most of the detection methods face low recognition accuracy among different BRB fault types, hence the accurate and reliable diagnosis of BRB faults is still a challenging task. Vibration signal contains substantial information reflecting the real working conditions of IM. Therefore, this paper presents an optimized application of cyclic modulation spectrum (CMS) based on vibration analysis to improve the BRB diagnosis, in which the spectral coherence and enhanced envelop spectrum (EES) is used to extract fault features more clearly. The STFT is a critical step in CMS analysis, therefore it is essential to determine the appropriate window function utilized in STFT according to BRB fault signatures in IM. Finally, the experimental analysis under a wide range of working conditions reveals the CMS with selected Kaiser (β = 3) window function can provide more accurate diagnostic information for different BRB severity compared to the popular envelope analysis.