Modified dictionary design of sparse decomposition for lamb wave structural health monitoring

Sparse decomposition is a promising approach for Lamb wave structural health monitoring. In Lamb wave sparse decomposition, designing an appropriate dictionary is of vital importance, which should be as accurate to actual wave packet as possible. Dispersion and multi-mode characteristics have been used for dictionary design. However, none of the previous research considered amplitude modulation in dictionary design, which limits the performance of sparse decomposition. To solve this issue, modified dictionary design for Lamb wave sparse decomposition is proposed in this study. First, frequency-dependent amplitude modulation curve is obtained based on the prior information of plate thickness, the type of mode, and the transducer size. Second, amplitude modulation is placed on the Lamb wave propagation model to redesign the dictionary. An experiment is given to validate the effectiveness of Lamb wave sparse decomposition with the modified dictionary design.