压力容器2025，Vol.42Issue(11)：66-74,9.DOI:10.3969/j.issn.1001-4837.2025.11.008
激光超声导波融合人工神经网络的复合材料缠绕压力容器损伤评估
A laser ultrasonic guided wave and artificial neural network-based approach for damage assessment in COPVs
摘要
Abstract
Composite overwrapped pressure vessels(COPVs)are widely used in industries such as chemical engineering and energy due to their superior properties of light weight,high strength and excellent corrosion resistance.However,during long-term service,inevitable damage initiation and propagation occur within the structure of COPVs,posing a significant threat to structural safety.Therefore,reliable non-destructive testing(NDT)techniques are urgently required to achieve quantitative evaluation of damage severity.This study proposes a two-dimensional regional stiffness inversion method combining laser ultrasonic guided waves with an artificial neural network(ANN),aiming to assess both interfacial debonding and impact-induced damage.For each spatial point,for training the ANN model,the semi-analytical finite element(SAFE)method is employed to generate wavenumber samples at different frequencies under predefined stiffness coefficients(i.e.,the outputs of the ANN).Subsequently,laser ultrasonic experiments are conducted to extract local wavenumber features,which are then fed into the trained model to realize the characterization of the two-dimensional distribution of the stiffness coefficient C11.The results indicate that at the impact point,the predicted stiffness reduction ratio of C11 is approximately 0.508(i.e.,the stiffness decreases to 50.8％of its original value),with a relative error of 4.85％compared with the theoretical results obtained via the SAFE method,demonstrating both the accuracy and feasibility of the proposed method.This approach provides a practical and feasible technical solution for the severity assessment of two-dimensional regional damage in COPVs.关键词
复合材料/激光超声/人工神经网络/局部波数/二维刚度Key words
composites/laser ultrasonics/artificial neural network/local wavenumber/two-dimensional stiffness分类
机械制造引用本文复制引用
杨乐辉,赵金玲,常乐,赵建平,程经纬..激光超声导波融合人工神经网络的复合材料缠绕压力容器损伤评估[J].压力容器,2025,42(11):66-74,9.基金项目
国家重点攻关项目(2023ZY01001) （2023ZY01001）
