Wavelet de‐noising techniques with power spectral density to vibration signal
Abstract
Purpose
Denoising of the vibration signal is crucial to identify a structure's damage. Based on noise frequency character, the “real” vibration signal can be gotten. The purpose of this paper is to propose a novel method for denoising a signal based on the wavelet transform.
Design/methodology/approach
The vibration signal with noise which can be collected by wireless network is decomposed by wavelet transform. In order to select optimal level of wavelet decomposition, based on noise's frequency, power spectral density is used. A soft thresholding method based on minimum mean‐variance is used for vibration signal de‐noising with Gaussian noise.
Findings
A novel method has been described in his paper. Based on the relationship between vibration signal's character and noise frequency, the way to get rid of noise is combined wavelet transform with power spectral density.
Originality/value
In order to select optimal level of wavelet decomposition, based on noise's frequency, power spectral density is used. A soft thresholding method based on minimum mean‐variance is used for vibration signal denoising with Gaussian noise.
Keywords
Citation
Chen, W., Wang, B., Zhan, H. and Zhou, L. (2013), "Wavelet de‐noising techniques with power spectral density to vibration signal", Kybernetes, Vol. 42 No. 4, pp. 604-613. https://doi.org/10.1108/K-10-2012-0076
Publisher
:Emerald Group Publishing Limited
Copyright © 2013, Emerald Group Publishing Limited