中文摘要：

薄板超声Lamb波层析成象研究中,走时是重要的信号特征,但多模式和频散特点给单一模式走时提取带来困难,一般信息提取技术对Lamb波分析较难有理想结果。该文提出利用HHT方法分析信号,先利用经验模式分解成本征模态函数,然后进行Hilbert变换分析边际谱并提取单一模式。与FFT、小波变换分析对比,HHT方法较好的反映了Lamb波信号多模式的随时间、频率分布。与小波变换提取走时数据结果相对比,HHT方法所得走时数据用来反演的图象更接近缺陷真实尺寸。

英文摘要：

Travel time is an important feature in lamb wave tomography. However, multiple modes and different dispersion properties make travel time extraction of single mode difficult. General feature extra- ction is unsuitable for lamb waves. HHT is used to analyze the ultrasonic lamb wave signal in this paper. lamb waves are first decomposed into intrinsic mode functions （IMF）. Each IMF is processed by Hilbert transform, and the marginal spectrum analyzed. It makes extraction of single modes easy. By comparing with FFr and wavelet transform, the result shows that HHT is an efficient and good method in showing the distribution of multiple modes with time and frequency. In contrast to wavelet transform, the lamb wave tomography generated with the travel time data by HHT is close to the real information of defect.