Vibration signal analysis has been widely used in the fault detection and condition monitoring of rotation machinery. But the practical signals are easily polluted by noises in their transmission process. The raw signals should be processed to reduce noise and improve the quality before further analyzing. In this paper an improved wavelet threshold denosing method for vibration signal processing is proposed. Firstly, a new threshold is developed based on the VisuShrink threshold. The effect of noise standard deviation and wavelet coefficient is retained, and the correlation of wavelet decomposition scale is considered. Then, a new threshold function is defined. The new algorithm is able to overcome the discontinuity in hard threshold denoising method and reduce the distortion caused by permanent bias of wavelet coefficient in soft threshold denoising method. At last five kinds of threshold principles and three kinds of threshold functions are compared in processing the same signal, which is simulated as the mechanical vibration signal added white noises. The results show that the improved threshold is superior to the traditional threshold principles and the new threshold function is more effective than soft and hard threshold function in improving SNR and decreasing RMSE.
[1]
Satish C. Sharma,et al.
Rolling element bearing fault diagnosis using wavelet transform
,
2011,
Neurocomputing.
[2]
C. Gargour,et al.
A short introduction to wavelets and their applications
,
2009,
IEEE Circuits and Systems Magazine.
[3]
Jing Yuan,et al.
Multiwavelet transform and its applications in mechanical fault diagnosis – A review
,
2014
.
[4]
Rui-Mei Zhao,et al.
Improved threshold denoising method based on wavelet transform
,
2015,
2015 7th International Conference on Modelling, Identification and Control (ICMIC).
[5]
Zhang Chunlei,et al.
Denoising algorithm based on wavelet adaptive threshold
,
2012
.