Abstract:

The detection of broken-rotor-bar faults based on common steady-state-analysis techniques, such as the fast Fourier transform (FFT), is known to be significantly dependent on the loading conditions of the induction motor. At light load, it is difficult to distinguish between healthy and faulty rotors because the characteristic broken-rotor-bar fault frequencies in the stator current are very close to the fundamental-frequency component and their amplitudes are small in comparison. As a result, detection of broken bar faults and classification of the fault severity under light load is very difficult. To overcome this problem, the analysis of the envelope of the transient starting-current waveform using the wavelet-transform has been investigated. The envelope extraction is used to remove the strong fundamental-frequency component, which would otherwise overshadow the characteristic differences between a healthy motor and a faulty motor with broken rotor bars. The wavelet-transform results are processed to develop a normalised parameter called the `wavelet indicator' which is sensitive to the presence of broken-rotor-bar faults. The results are verified using tests on machines with a varying number of broken bars, as well as partially broken rotor bars, over a wide range of loading conditions. The effects of initial rotor position and supply imbalance are also investigated.