大型装备关键结构件常承受交变或冲击载荷，表面易萌生疲劳裂纹，目前针对疲劳裂纹的快速检测需求巨大。常规无损检测方法在疲劳裂纹快速识别、量化、重构等方面存在诸多技术瓶颈，项目综合考虑涡流热像和振动热像技术的优势和局限，提出涡流-振动复合激励红外热像技术，拟以薄壳车体和钢轨两类金属结构件中的垂直疲劳裂纹和滚动接触疲劳裂纹为研究对象，构建三维解析和数值模型，探寻电磁-机械-热多物理场耦合下裂纹区域的生/散热机理；优化激励信号和装置，实现复合激励下裂纹热响应增强，并基于卷积-循环神经网络和热张量分解方法，实现裂纹特征提取；研究反向热扩散稳定性问题，解决裂纹形貌/热源分布的三维重构难题；最终，以检出概率为工具，进行复合激励参数下裂纹可检测性定量评估。本研究紧贴在役装备关键结构快速检测的应用需求，其研究成果将丰富无损检测领域针对疲劳类裂纹识别量化的知识储备，具有重要的理论意义和应用价值。

Under long time alternating or impact loads, surface-breaking fatigue cracks may initiate in key structural components of engineering equipment. Fast testing of fatigue cracks is of urgent demands. To address the challenges of conventional nondestructive testing methods in identification, quantification, and reconstruction of fatigue cracks, this project proposes the use of induction and vibration dual-excitation infrared thermography by considering the advantages and limitations of induction thermography as well as vibrothermography. Focusing on testing two types of fatigue cracks, i.e., vertical fatigue cracks in thin shells and rolling contact fatigue cracks in rails, this work carries out the following investigations: (1) implementations of three-dimensional analytical and numerical models for investigating the mechanisms of thermal heating and diffusion in the crack vicinity under induction and vibration excitations; (2) enhancement of crack thermal responses based on optimizing excitation signals and configurations, and crack feature extraction based on convolution-recurrent neural network and thermal tensor decomposition; (3) crack shape and heat source reconstruction based on regulation of inverse thermal diffusion; (4) probability of detection analysis for crack detectability evaluation. This project is closely related to the testing of key structural components from in-service equipment. Research results aim to enrich the knowledge base of non-destructive testing for identification and quantification of fatigue cracks, which have great theoretical and application values.