04.1 - Material Issues, Fatigue and Damage Tolerance
A DEEP-LEARNING-BASED METHOD FOR DAMAGE IDENTIFICATION OF COMPOSITE LAMINATES
Y. Yang¹, B. Wang¹, S. Lv¹, Y. Zhou, Peking University, China; L. Wang¹, G. Liu¹; ¹Aircraft Strength Research Institute of ChinaAffiliation, China
Damage detection based on guided wave technology is especially suitable for long-term monitoring of aircraft composite structures. However, current methods, no matter traditional physical-model-based or machine-learning-based approaches, require preprocessing for extracting hand-crafted features in advance. This leads to great limitations as they greatly depend on the knowledge and experience of individual experts. We propose a deep learning method TDCNN, which effectively overcomes previous drawbacks and exhibits great performance through verification.