Guided-wave quick sparse decomposition approach based on novel peak-frequency modulation dictionary for efficient composite plate damage inspection

• Published in: Applied acoustics Vol. 223; p. 110081
• Main Authors: Hong, Xiaobin, Luo, Zewen, Zhang, Bin, Jin, Gang
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 05.07.2024
• Subjects: Composite; Matching pursuit; Modulation dictionary; Structural health monitoring; Ultrasonic guided wave; Modulation dictionary; Structural health monitoring; Composite; Ultrasonic guided wave; Matching pursuit
• ISSN: 0003-682X; 1872-910X
• DOI: 10.1016/j.apacoust.2024.110081

•A novel peak-frequency modulation over-complete dictionary considering peak number, signal frequency and attenuation is constructed.•Peak-based quick matching pursuit algorithm (PQMP) is designed to realize the efficient separation, extraction and recognition of the guided wave mode.•Multi-sensor elliptic probability imaging algorithm is improved to visually obtain the damage location.•A novel imaging accuracy index (IA) is constructed to evaluates the imaging performance. High performance carbon fiber reinforced plastics structure is a strategic material for mechanical and engineering equipments, and its whole structural health monitoring is particularly important. Ultrasonic guided wave monitoring technology is a promising means to monitor this composite structure, but the separation and extraction of ultrasonic guided wave modes are still a challenge. To address this problem, a novel guided-wave quick sparse decomposition approach is proposed to separate and extract the guided wave mode and assess health state of the composite plate. Firstly, a novel overcomplete peak-frequency modulation dictionary considering the peak number, the signal frequency and the attenuation of mode is established to adapt to different dispersion characteristics of guided wave signals. Then, peak-based quick matching pursuit algorithm (PQMP) is designed to realize the efficient separation, extraction and recognition of the guided wave mode. Finally, multi-sensor elliptic probability imaging algorithm is improved to visually obtain the damage location. Experimental results show that the proposed method can efficiently and accurately separate the guided wave mode information. The running time is only 1/11 of the MP method and 1/9 of the OMP method, and the matching accuracy is about 0.9. And the proposed method can diagnosis the damage position of composite plates clearly. The imaging accuracy index(IA) of the proposed method is over 0.9.