Adaptive Wavelet Threshold Ｒolling Bearing Fault Vibration Signal Denoising Method

Abstract

For the noise reduction processing of mechanical bearing fault vibration signals，the existing wavelet threshold noise reduction method can not meet the requirements of extracting weak vibration signals of bearings. Aiming at the problems of hard threshold and soft threshold wavelet denoising methods with discontinuity and constant deviation and poor flexibility of existing threshold functions，an adaptive wavelet threshold function is proposed to denoise the rolling bearing fault vibration signal. In this paper，the bearing data of Western Ｒeserve University is used to simulate and analyze the actual project. Compared the noise reduction images of the frequency domain signal refinement spectrum with different threshold functions，the adaptive wavelet threshold function can better filter out the noise and extract the weak vibration signal.