高级检索

    基于压缩感知的Lamb波信号成分分离

    Lamb wave signal component separation based on compressed sensing

    • 摘要: 损伤检测中常用的超声导波为Lamb波,其传播距离长,灵敏度高,但多模态及频散效应降低了其在时间和空间上的分辨率。提出了一种基于压缩感知的Lamb波成像处理方法,该方法需要构造精确的字典来匹配损伤反射Lamb波信号。利用MATLAB软件模拟得到不同行进距离的单模态信号,通过该信号来建立单模态字典;采用压缩感知得到系数矩阵,选择出单模态字典中最匹配损伤反射信号的原子,并用其表征原始信号以完成模态分离。建立了非频散字典,最终得到单模态、非频散的信号。

       

      Abstract: Lamb wave, as a commonly used guided wave in structure detection, is characteristic of long propagation distance and high capability, but its multi-mode and dispersion effect reduces its temporal and spatial resolutions for its application, to some extent. A Lamb wave imaging processing method based on compressed sensing is proposed, which needs to construct a reasonable dictionary to match the reflected Lamb wave signal. Firstly, MATLAB numerical simulation was used to obtain single-mode signals with different travel distances to build a single-mode modal dictionary; then the single-mode modality was selected by compressing the matrix. Similarly, a non-dispersion dictionary was built up, and a single-mode signal with no dispersion was obtained.

       

    /

    返回文章
    返回