The de-noising approach of plain bearing of AE signal based on multi-level and multi-position sparse representations
ZHANG Junning1，ZHANG Peilin1，CHEN Yanlong1，WANG Huaiguang1, SUN Yezun2 ，YANG Wangcan1
1 Department 7st Ordnance Engineering College, Shijiazhuang 050003, China
2 Beijing Military Delegation Office Located at Factory 247,Taiyuan 030009,China
Abstract：Acoustic Emission (AE) signal is blurred seriously by noise, which limits the denoising capacity of K-means Singular Value Decomposition(K-SVD) dictionary algorithm. Based on this, the moving ruler based K-SVD (KMR-SVD) dictionary for quickly Denoising AE signal is proposed considering the characteristics of AE signal in this paper. Firstly, the multi-level and multi-position sparse characteristics is obtained by using the moving ruler strategy. Thus it solves the issue that the K-SVD ignores the hidden information among atoms. Secondly,the computing speed of dictionary algorithm is improved by applying Absolute Grey Relational Degree of B-mode(AGRDB) to reduce the redunancy of atoms. Therefore, compared with the traditional K-SVD algorithm, this algorithm can achieve better denoising performance. Moreover, it can preserve information of local feature as well. The experimental results show the change of the friction bearings state, which validates the effectiveness of the algorithm.
张峻宁1，张培林1，陈彦龙1，孙也尊2，杨望灿1，. 基于多层多位置稀疏的滑动轴承AE信号降噪[J]. 振动与冲击, 2016, 35(19): 107-112.
ZHANG Junning1，ZHANG Peilin1，CHEN Yanlong1，WANG Huaiguang1, SUN Yezun2,YANG Wangcan1. The de-noising approach of plain bearing of AE signal based on multi-level and multi-position sparse representations. JOURNAL OF VIBRATION AND SHOCK, 2016, 35(19): 107-112.
