基于稀疏表示的兰姆波模态分离方法研究

针对超声阵列检测过程中兰姆波的频散和多模态特性带来的信号分析困难问题,提出了一种基于稀疏表示的兰姆波模态分离方法,并将其应用于金属板结构兰姆波模态分离中。根据S;和A;模态兰姆波在时域上的波形差异,研究S;、A;单模态字典和复合字典的构造方法,将待测信号在复合字典上进行稀疏分解,采用带相干性准则的最小角回归算法求解稀疏系数,从多模态信号中提取S;、A;单模态信号。将所提出的方法应用到仿真和实验信号分析中,结果表明,该方法可以很好地实现检测信号中单模态兰姆波信号的提取。研究工作有助于解决金属板结构中多模态兰姆波的模态分离问题。

Method of Lamb Wave Mode Separation Based on Sparse Representation

In view of the problem of signal analysis caused by the dispersion and multi-mode characteristics of Lamb wave in ultrasonic array detection,a method of Lamb wave mode separation methods based on sparse representation is proposed and applied to the Lamb wave mode separation of metal plate structure.According to the waveform difference between S₀ and A₀ mode Lamb waves in time domain,the construction methods of S₀ and A₀ single mode dictionary and composite dictionary are studied.The signal to be measured is decomposed on the composite dictionary.The least angle regression algorithm with coherence criterion is used to solve the sparse coefficient,and S₀ and A₀ single mode signals are extracted from the multi-mode signals.The proposed method is applied to simulation and experimental signal analysis,and the results show that the method can well achieve the extraction of single-mode Lamb wave signal in the detection signal.The research work is helpful to solve the mode separation problem of multi-mode Lamb wave in metal plate structure.