Feature Extraction of Acoustic Emission Signals for Rotor Fault Based on Daubechies Wavelet Analysis

Abstract

The failure of rotating machine is often occurred by the rotor rubbing and the acoustic emission will be created by the rotor rubbing. The acoustic emission (AE) signal is the transient and non-stationary signal. Wavelet Transform has outstanding advantage revealing the character of this signal in time-frequency. With the increasing of the rotor speed, the noise increases a lot and the wavelet coefficients of AE signal change to high-scale. The db10 was applied in AE signal processing, and the modulus maxima method was applied to reconstruct of the signal. The result shows that the AE signal of rotor rubbing could be separated from noise by Daubechies Wavelet