Rotating machine fault diagnosis based on correlation filtering and reconstructed Morlet wavelet power spectral entropy

The main method of fault diagnosis is to determine the fault location by extracting the fault frequency after the original signal is de-noised, but in many cases it is not enough to judge the fault location only by the fault frequency. In this paper, the correlation filtering method are combined to Morlet wavelet and power spectral entropy to judge the faults of different degrees at the same frequency based on extracting the fault frequency, and the validity of this method is verified by testing the actual bearing fault data. First, a set of wavelet groups with a high correlation coefficient and a maximum correlation coefficient of the original signal were extracted through correlation filtering method based on Morlet wavelet. Then the specific wavelet in the wavelet group was linearly superimposed. finally, the power spectral entropy of the superimposed wavelet was calculated. The power spectral entropy and periodic parameters are used to judge the bearing fault location and its severity. Result indicate that the proposed method is able to distinguish between normal and fault signals and provide a basis for judging the degree of failure.