A synchrosqueezing transform based method for rotor fault diagnosis of squirrel-cage induction motors

For the diagnosis of rotor faults in squirrel-cage induction motors (SCIM), the time-frequency (TF) analysis of the stator current signal plays an important role in many effective diagnosis methods, especially in those aimed at the run-up and frequency conversion adjusting speed periods. However, the feature that two sideband frequencies close to fundamental frequency of rotor fault signal usually cannot be directly extracted from the blurred time-frequency representation (TFR) caused by bad resolution of traditional TFR, such as WT and STFT. To solve this problem, a new method based on the synchrosqueezing transform (SST) is studied to detect and diagnose rotor faults. The stator current signal is first mapped into the wavelet basis to facilitate synchrosqueezing of the TF picture in the direction of frequency-scale domain. The clear instantaneous frequency profiles with concentrated TFR are presented. The rotor fault, if any, can then be detected by observing the presence of fundamental frequency and sideband components in the TFR. The proposed method is evaluated by using both simulated data and experimental stator current signals. All results show that the SST-based method outperforms the existing methods for SCIM rotor fault diagnosis.