Multiple bandwidth wavelet analysis for fault diagnosis eigenvalue in squirrel-cage induction motor

The fault of rotor broken bars in three-phases induction motor is diagnosed by using the multiple bandwidth wavelet analysis in this paper. Daubechies wavelet is selected as the mother function, and the wavelet coefficient is obtained from multiple bandwidth wavelet analysis of the fault current signal of the induction motor. The wavelet reconstruction of each frequency branch is then processed, which is the time signal of the current in each frequency branch. In this paper the energy eigenvalue in each frequency branch is defined as the criterion to diagnose the rotor fault. The rotor fault of the induction motor is clearly identified not only from the distinguishing eigenvalue in the low frequency branch, but also from the eigenvalue in some high frequency branch, compared with the no fault condition. Therefore the multiple bandwidth wavelet analysis provides a more sensitive means to diagnose faults than does the Fourier transform.