基于性能退化信息的复杂系统单机状态监测与动态可靠性分析

With the technical development of sensor and intelligent monitoring, traditional population assessment and static analysis can no longer satisfy the practical need from current reliability engineering. Full service life cycle dynamic reliability analysis and remaining useful life prediction for individual complex system has become a new challenge in fields of aeronautics and astronautics. .Nowadays, most systems show high reliability and long life properties, and then it is difficult to obtain sufficient failure life data through traditional life tests and accelerated life tests. Driven by monitoring performance data, the current study investigates the key issues in individual complex system degradation modeling, operating status identification and dynamic reliability analysis adopting both theoretical and experimental means. First, a Gaussian process degradation model is proposed via population and individual life information fusion. Second, an online changepoint detection method is established for real-time operating status monitoring for an individual system. Finally, a multi-stage degradation model is built and an effective online model updating mechanism is proposed based on Bayesian and Kalman filter theory. Consequently, life-cycle online reliability analysis and useful residual life prediction for an individual complex system can be achieved. The research achievement will be of scientific significance and practical value for the development of dynamic reliability analysis and maintenance decision-making for individual complex systems.

新型传感器技术突破和智能监测水平提升，使当前可靠性工程不再满足于传统的总体评估和静态分析，复杂系统单机全服役周期的动态可靠性分析和寿命监控已成为航空航天、武器装备等领域的新方向和新挑战。.针对高可靠、长寿命等共性特点使多数系统难以通过传统的寿命试验和加速寿命试验获得足够失效寿命数据的问题，本项目以性能监测数据为驱动，采用理论和试验相结合的手段，对单机性能退化建模、运行状态识别、动态可靠性分析等关键问题开展研究。具体包括：基于总体与个体寿命信息融合，构建更为合理的高斯过程单机性能退化模型；建立单机运行状态变点的在线识别方法，实现单机服役期间运行状态的实时监测；建立多阶段退化模型，并基于Bayes和Kalman滤波理论形成模型的在线更新机制，实现面向全寿命周期的单机动态可靠性分析。研究成果对复杂系统单机动态可靠性分析理论的完善、可靠性评估和维修决策制定技术的工程应用具有重要的科学意义和实用价值

在航空航天、武器装备等领域，复杂系统单机全服役周期的动态可靠性分析和寿命监控已成为备受关注的新方向和新挑战。本项目以单机性能监测数据为驱动，采用理论与试验相结合的手段，对总体和个体退化寿命信息的融合建模、系统运行状态变点在线识别以及多阶段退化建模与在线分析等关键问题展开深入研究。.通过揭示非线性退化首达时寿命分布机理和总体与单机的寿命信息融合机制，构建了单机非线性退化过程模型；针对运行环境切换损伤特征改变的现象，基于极大似然理论建立了统计检验方法，实现变点的在线识别；考虑单机不同运行阶段的失效特征变化，构建了多阶段非线性随机过程模型；在此基础上，建立了单机性能实时监测与信息更新方法，实现单机动态可靠性分析和剩余寿命预测。本项目研究对可靠性理论的完善和装备长期服役可靠性建模、验证与健康管理具有理论价值和实践意义。