基于导波原位检测的复合材料疲劳表征与寿命预测研究

随着复合材料在先进飞行器结构中占比的逐渐增加，复合材料在服役过程中力学性能的变化对飞行器整体的安全至关重要。为了实现基于导波原位检测的飞行器复合材料整体部件疲劳评估和寿命预测，首先，从宏观和细观的角度研究复合材料疲劳损伤演化规律；在此基础上，通过分析导波波场信息，探究导波相速度、模态能量比等特征在表征复合材料疲劳方面的潜力；其次，从复合材料损伤机理出发，建立导波相速度与疲劳损伤累积的演化模型；然后，构建深度学习框架，以数据驱动的方式从导波波场中提取疲劳演化特征；最后，提出基于贝叶斯模型平均方法的疲劳演化模型，对复合材料剩余疲劳寿命进行预测。结果表明：通过提取和分析导波特征信息，可以准确地对复合材料疲劳状态进行表征，结合贝叶斯模型平均方法和置信区间准则，实现了在试件疲劳破坏之前的剩余寿命预测。

Research on fatigue characterization and life prediction of composites based on guided wave in-situ detection

As composite materials are playing a more and more important role in advanced aircraft structures, the change of mechanical properties of composites during service is very important to the overall safety of the aircraft. In order to achieve the goal of fatigue evaluation and life prediction of composite components of aircraft based on guided wave in-situ detection. This paper carries out research on three aspects: the degradation law of structural mechanical properties, the influence mechanism of fatigue accumulation on guided wave propagation, and methods on fatigue characterization and life prediction. First of all, the fatigue evolution law of composite materials is studied from the perspectives of macroscopic phenomenology and microscopic physics. Then, the potential of guided wave phase velocity and mode conversion phenomenon for fatigue characterization is discussed through analyzing the guided wave field. At the same time, a deep learning framework is constructed to extract fatigue evolution features from the guided wave field in a data-driven manner. Finally, a fatigue evolution model based on the Bayesian model averaging method is proposed to predict the residual fatigue life of the composite specimen. Results show that: by extracting and analyzing the guided wave propagating features, the fatigue state of composite materials can be accurately characterized. Combining the Bayesian model averaging method and the confidence interval criterion, the goal of residual life prediction before specimen fatigue failure can be achieved.