Research On The Detection Methods Of Induction Motor Fault Based On Electrical And Mechanical Signal

The detection methods of rotor bar breaking fault and bearing fault in induction motors are investigated and realized in this paper. On the basis of the summary of the advantages and disadvantages of the classical methods to detect the rotor bar breaking fault, two detection methods, which are based on Hilbert transform envelope demodulation, Hilbert transform digital filter, self-adaptive filter, continuous subdivision Fourier Transform, rotor slot harmonics based slip estimation techniques and the modern power spectrum estimation method, Multiple Signal Classification Algorithm, are proposed. This paper emphasizes to research the detection methods of the bearing fault based on the vibration signal. A two-parameter simple detection method based on the time domain parameter, and a precision method based on delayed correlation-envelope technique, Hilbert-Huang transform and Multiple Signal Classification Algorithm are reviewed and realized. Fault detection instances in laboratory demonstrate that the above-mentioned detection methods are valid and effective.