Singularity detection and feature extraction based on analytic wavelet transform

A singularity detection and denoising method based on analytic wavelet transform(AWT) and signal reconstruction was proposed to improve the accuracy of signal singularity detection and the efficiency of fault diagnosis.According to the difference of propagation characteristics of wavelet transform modulus maximum(WTMM) of signal and noise along the scale direction,signal denoising and fault feature extraction were realized.Singularity detection and denoising based on AWT was applied to the vibration signals of running machines.Signals sampled under several conditions in a main reducer performance test bed were analyzed,and the fault diagnosis of the main reducer was conducted by analytic WTMM and real WTMM respectively.Experimental results show that the singularity detection using the modulus maximum of an analytic wavelet is better than that of a real wavelet,and that the fault feature can be distinguished more obviously and accurately.