强噪声非稳态多场耦合多损伤失效机理与微损伤智能检测诊断研究

In a complex environment, it's difficult to reveal multi-field coupling damage mechanism and to real-time detect and monitor multidimensional-damage. In the project, the fatigue fracture theory, the 3D damage tolerance theory, modern signal processing theory and modern testing technology will be perfectly combined to research the occurring mechanism of fine scale damage, the law of damage evolution, the micro-damage (less than 0.5mm) monitoring and the feature extraction mechanism of weak damage signal in the multi-physics coupling action，in the basis，establish the damage failure prediction model and the life evaluation system, and propose a practical CI (Computational Intelligence) methods that will be used to on-line detect and diagnose the running state of complex mechanical system under excitation of random condition, and develop a security monitoring system based on manifold learning that can accomplish the monitoring of system structure running state. At the same time, a on-line detecting communication platform that can solve the difficult problem of diagnosing micro-damage and multi-fault of high-speed running complex mechanical systems (such as aerospace structures, high-speed train structure) in the action of strong noise non-steady random excitation and multi-field coupling to overcome the drawbacks of poor diagnosis accuracy and monitoring, and micro-damage recognition accuracy will reach 0.2mm, the reaction time of online detecting will reach 1 second, and achieve the real-time detection, diagnosis and monitoring in whole life process to prevent the occurrence of major unforeseen accidents, and a new theory and method will be offered for the damage failure prediction forecasting and intelligent control, intelligent fault diagnosis and control of complex mechanical systems.

针对复杂环境下多场耦合损伤机理难以揭示和多微损伤难以实时检测监控难题，将疲劳断裂理论、三维损伤容限理论、现代信号处理新理论和现代测试技术有机结合起来，研究多场耦合作用下多损伤发生机理、细观尺度下微损伤产生扩展演化规律、微损伤（小于0.5mm）的可监测性、检测与微弱损伤信号特征提取机理，从而建立复杂系统的损伤失效预测评估模型和寿命评价体系，并提出一种实用的随机环境激励与多物理场作用下复杂系统运行的在线诊断与监控的计算智能CI方法，研制基于CI的流行学习式可实现系统结构运行的MEMS安全监控系统,最终建立系统在线检测与监控的理论和方法，解决强噪声非稳态随机激励多场耦合下高速运行复杂系统（如航空航天结构、高速列车结构）微损伤和多故障检测诊断难题，克服以往诊断中识别精度低、可监测性差的弊端，使得微损伤识别精度达到0.2mm、在线检测反应时间达到1秒,从而实现对系统整个全寿命过程实时检测诊断与监控

为了揭示复杂环境下多场耦合损伤机理和多微损伤实时检测技术，将疲劳断裂理论和现代测试技术有效的结合起来，研究多场耦合作用下多损伤发生机理和细观尺度下微损伤产生扩展演化规律的可监测性与微弱损伤信号特征提取机理，建立复杂系统的损伤失效预测模型。提出一种实用的随机环境激励与多物理场作用下复杂系统运行的在线诊断与监控计算方法，建立了系统在线检测与监控的理论和方法，解决强噪声随机激励多场耦合下复杂系统微损伤和多故障检测诊断难题，提高微损伤识别精度并实现对系统全寿命过程实时检测诊断。