基于经验模态分解去噪的粗晶材料超声检测  

Ultrasonic Testing of Coarse-grained Materials Based on EMD Denoising Method

作　　者：李秋锋[1,2] 黄攀[1] 施倩[1] 陈果[1] 陈振华[1] 

机构地区：[1]南昌航空大学无损检测技术教育部重点实验室,江西南昌330063 [2]近代声学教育部重点实验室(南京大学),江苏南京210093

出　　处：《应用基础与工程科学学报》2014年第3期566-573,共8页Journal of Basic Science and Engineering

基　　金：国家自然科学基金项目(11264032;11104129);航空科学基金项目(2011ZE56006);江西省自然科学基金项目(20122BAB201024);南昌航空大学研究生创新基金项目(YC2012012);江西省研究生教育创新基地资助项目;江西省教育厅科学技术研究项目(GJJ14530)

摘　　要：粗晶材料超声检测中,结构噪声严重降低了检测信号的信噪比,缺陷反射难以识别.为了增强检测信号信噪比,提高粗晶材料超声检测的可靠性,采用经验模态分解(EMD)技术对检测信号进行去噪处理,通过3次样条插值形成波形包络,并利用信号的特征时间尺度将非线性、非平稳检测信号自适应的分解成多个本征模态函数(IMF)之和,从而获得信号高阶成份和趋势.利用EMD的这种特性对低信噪比模拟信号进行处理,并将处理结果与小波去噪结果进行对比,信噪比获得更大提高.通过对粗晶材料实测信号进行去噪实验,结果表明EMD去噪具有更强的自适应能力,且需知的原信号先验信息更少.In ultrasonic testing of coarse-grained materials, Signal to Noise Ratio （SNR） of detection signals was reduced seriously for the structure noise, and echoes from defects were difficult to be identified. Empirical Mode Decomposition （EMD）was introduced to process the testing signal in order to improve the SNR and the reliability in ultrasonic testing of coarse-grained materials. Signal envelope could be formed by using cubic spline interpolation, and nonlinear and non- stationary signal could be decomposed self-adaptive into the sum of some Intrinsic Mode Functions （IMF） by using characteristic time scale of the signals, and the higher order components and tendency of the original signals could be obtained. The denoising experiment with low SNR simulated signal were achieved according to the feature of EMD, and SNR was enhanced more by comparison with the wavelet analysis method. And the detection signal collected from coarse-grained materials was used to achieve experiment, and the experimental results show that the EMD has better adaptive ability in decomposing noise-polluted signals and less empirical information is required in the denoising process.

关 键 词：粗晶材料 超声检测 经验模态分解 信噪比 

分 类 号：TB551[理学—物理]