基于形态滤波与EEMD组合方法的声发射去噪  

De-noising of acoustic emission signals based on the combination of morphological filtering and EEMD

作　　者：席剑辉[1] 许廿 

机构地区：[1]沈阳航空航天大学,辽宁沈阳110036

出　　处：《制造技术与机床》2016年第12期70-75,共6页Manufacturing Technology & Machine Tool

基　　金：国家自然科学基金青年基金资助项目(60804025);辽宁省教育厅科学技术研究项目(L2014069)及(L2013070);沈阳市科技创新团队项目(src201204)

摘　　要：针对声发射信号传递和采集过程中干扰噪声多态性、随机性等问题,研究形态滤波与集合平均经验模态分解(EEMD)的组合方法,从空间域和时频域不同角度对信号进行降噪处理。首先采用加权级联形态滤波,对闭开和开闭运算进行适当搭配,通过权值调整有效滤除声发射信号的尖峰脉冲干扰;然后应用EEMD对处理后的信号进行分解,依据相关系数去除虚假分量并重构信号,提取信号特征信息。对模拟声发射信号和实测声发射信号的仿真结果说明了此方法的有效性。Aimed at the problems in the process of acoustic emission signal transmission and acquisition, such as the diversity and the randomness of the interference noise, a combination methodology of the morphological filtering and the ensemble average empirical mode decomposition （EEMD） was studied, then the signal de - noising was carried out from different angles of spatial domain and time frequency domain. Firstly, use the weighted cascaded morphological filtering, the close -open and open -close operations of which were appropriately matched, to effectively filter the peak pulse interference of acoustic emission signal by the weight adjustment. Then the EEMD method was used to decompose the processed signal. According to the correlation coefficient method, the false components could be removed and the signal could be reconstructed for extracting signal feature information. Simulation results from both the simulated and the practically measured acoustic emission signals were presented to illustrate the effectiveness of the proposed method.

关 键 词：声发射 形态滤波 EEMD 去噪 

分 类 号：TH806[机械工程—仪器科学与技术] TH113[机械工程—精密仪器及机械]