A Lightweight CNN-RPA Network for Guided Wave Based Real-Time Damage Detection in Composite Structure

Ultrasonic guided wave (UGW)-based damage detection is considered one of the most promising structural health monitoring (SHM) technologies for assessing the safety and integrity of composite structures. However, achieving accurate and effective real-time damage detection remains a challenge. To address this issue, a novel UGW-based damage detection approach is proposed for real-time damage localization in composite plates. In the proposed approach, considering the expensive calculation of multi-path UGW signals, an efficient UGW signal compression method is designed based on the improved differential-driven piecewise aggregate approximation (IDPAA) algorithm to enhance calculation efficiency. Next, by incorporating the residual pixel-level attention (RPA) mechanism, the proposed lightweight CNN-RPA network can detect damage with high accuracy and low computational complexity. Notably, the performance of the proposed lightweight network is verified using a real-world dataset. Experimental results demonstrate that the proposed damage detection approach delivers exceptional performance in locating damage, surpassing mainstream end-to-end damage detection methods in term of accuracy and efficiency.