Diagnosis of Multiple Rotor Bar Faults of Squirrel Cage Induction Motor (SCIM) Using Rational Dilation Wavelet Transforms

Abstract

Vibration is the vital feature of the motor to diagnose various faults. The existing methods have many limitations. This work proposes a non-invasive Software Phase Locked Loop (SPLL) technique to diagnose the multiple rotor faults. A high-frequency (10.525GHz) microwave signal is projected on the motor and the reflected signal is captured by the handheld ultra-wide band (UWB) radar. Hardware filter is replaced by a software phase locked loop filter to remove high-frequency noise content in the signal. The output of the SPLL is analysed with Rational Dilation Wavelet Transforms (RDWT) and compared the change in the signal energy of the wavelet sub-band 6 for the multiple rotor bar faults and sub-bands 6 and 7 for the combination of rotor bar and bearing faults, respectively. The signal energy at sub-band 6 increased from 2.19% without fault to 3.99%, 6.03% and 13.66% with an increase in the number of rotor bars with faults. The variation of the experimental values of the signal energy for rotor bar faults compared with the theoretical values found that the error is reduced and accuracy is increased with the number of faults. For simultaneous rotor bar and bearing faults, the signal energy at sub-bands 6 and 7 also increases. The proposed SPLL-based method that is able to diagnose the individual faults and combination of the faults and has advantages such as simple, cost-effective and non-invasive technique compared with the existing methods.