基于多应力时间效应的引信控制系统故障预测及健康管理方法研究

The prognostic and health management issue of long-term storage equipment has always been a hot research topic which needs to be solved urgently in the field of PHM technology,this issue has received much attention from scholars around the world. Most of the existing methods are based on single environmental stress and single fault parameters, but the long-term storage issue has itsparticular characteristicwhich makes traditional methods difficult to apply. Especially for the special application background that can not be electronic measured or can only have short-term and limited-point electronic measurement, the traditional method leads to the problem of low prognosticaccuracy. In view of this, this study will discuss the fault diagnosis and lifetime prediction problems during the equipment storage period. In order to provide new methods and technical means for the aboveurgent topic, the typical fuze control circuit system with "long-term storage, one-time use"characteristics will be used as the research object here. In this proposal, this research will start from the multi-stress and time effects modeling problem, then, focus on the following orientation: parameters extraction methodology study considering physics-electronics based multi-stress effect coupling,the diagnosis method study on multivariate correlation model with multi-stress effect coupling, and the real-time remaining useful performance prediction modeling study when the equipment is during storage period. The key theoretical problems proposed in this subject have important theoretical and practical values for expanding the PHM technical field and solving the equipment health management problems in itslong-term storage environment.In addition, this research will further promote the development of PHMtechnology to the applicationorientation.

长时间贮存装备的故障诊断及健康管理一直是PHM技术领域亟待深入解决的热点问题，倍受世界各国学者的高度关注。现有方法多是从单一环境应力以及单一故障参量出发进行研究，而长时间贮存的特殊性使得传统方法还难以适用。特别是对于不能加电或仅能短时有限点加电测量的特殊应用背景，传统方法导致的诊断精度或预测精度低的问题更为突出。有鉴于此，课题将针对贮存状态下的装备诊断及寿命预计问题，以引信控制系统这一典型的“长期贮存、一次使用”电路为背景，从多应力和时间效应出发，通过对基于物理-电参量关联分析的参数获取、多应力多参量关联诊断方法、贮存环境下的动态剩余寿命预测方法的研究，为贮存环境下的装备诊断及预测提供新的方法和技术手段。本项目解决的关键理论问题，对于拓展PHM的技术领域，解决装备长期贮存环境下的健康管理问题具有重要的理论意义和应用价值，将进一步推动PHM预测技术向实用化方向发展。

引信控制电路系统是一种复杂的数模混合电子系统，系统内部数模混合芯片不断增多、结构日趋复杂、参数间的相互关联耦合不断增强，故障率也随之增加，由此引发的测试代价高、诊断难的问题愈发突出。现有方法多是从单一环境应力以及单一故障参量出发进行研究，而长时间贮存的特殊性使得传统方法还难以适用。特别是对于不能加电或仅能短时有限点加电测量的特殊应用背景，传统方法导致的诊断精度或预测精度低的问题更为突出。有鉴于此，课题将针对贮存状态下的装备诊断及寿命预计问题，以引信控制系统这一典型的“长期贮存、一次使用”电路为背景，从多应力和时间效应出发，开展数模混合电路有限特征的参数获取和健康表征、多应力多参量关联的故障诊断方法、贮存环境下基于等效系统分析的剩余寿命预测方法研究。经过三年的研究，本项目提出了一套基于多应力时间累积效应的电子系统PHM预测方法，在关联参数获取及健康表征、多信号关联建模、多应力故障诊断、多参量时间效应故障预测及寿命估计等方面取得了一定突破，通过融合物理失效特性，建立了多应力时间累积效应下的多种预测模型，实现了非特定应力下的电路性能预测能力，提升了对复杂电路实时可靠性预测性能，解决了数模混合装备贮存环境下的维护保障问题。项目组共发表高水平学术论文21篇（SCI论文11篇，中文期刊EI论文5篇，高水平国际会议EI论文5篇），新申请国家发明专利25项（其中，已授权4项），共培养博士生4名，硕士生16名。研究成果可直接应用于复杂电子装备的PHM系统中，可加强我国PHM技术在复杂电子系统中的应用能力，为提高我国军、民用电子装备的维护保障与可靠性提供技术支撑，从而进一步推动该项技术向实用化方向发展。