基于多阶段随机退化过程建模的剩余寿命预测与健康管理方法研究

Accurate prediction of remaining useful life (RUL) and reasonable health management are the key technologies to ensure safe and reliable operation of equipment. The existing researches mainly focus on the degradation modeling and RUL predicting of the continuous stochastic degradation process under a single operating environment and operating condition. However, due to environment changing and operating condition switching, equipment degradation processes often exhibit multi-stage feature, which makes existing methods unsuitable. In view of this, the project concentrates on the methods of RUL predicting and health management based on multi-stage stochastic degradation process modeling. In addition, this project focuses on the description of residual degradation process, solution of RUL’s distribution under multi-stage influence, and optimal decision making of maintenance and spare parts supply relying on forecast information. It mainly includes: 1) degradation modeling and RUL prediction for multi-stage deteriorating process with random mutation; 2) RUL prediction based on the multi-stage nonlinear degradation modeling; 3) joint decision-making method for predictive maintenance and spare parts supply. Finally, application verification will be carried out based on the practical data of some core devices equipped on missile control system. This project is expected to enrich the basic theory of RUL prediction for stochastic degradation equipment, and provide valuably theoretical and technical support for RUL prediction and health management of multi-stage stochastic degradation equipment.

准确预测剩余寿命并在此基础上制定合理的健康管理措施，是保障设备安全可靠运行的关键技术。现有研究主要集中在单一运行环境、单一工况下设备连续随机退化过程的建模及剩余寿命预测问题。然而，受环境改变和工况切换的影响，设备退化过程常常呈现出多阶段特征，已有方法难以适用。鉴于此，本项目拟开展基于多阶段随机退化过程建模的剩余寿命预测与健康管理方法研究，重点解决多阶段影响下随机退化过程的表征与剩余寿命分布的求解、预测信息下维护与备件供给的最优决策等关键问题，主要包括：1)带随机突变的多阶段退化建模与剩余寿命预测；2)基于多阶段非线性随机退化过程建模的剩余寿命预测；3)基于剩余寿命信息的预测维护与备件供给联合决策方法。最后，依托某型导弹控制系统关键设备实际数据，开展应用验证。通过本项目，有望丰富随机退化设备的剩余寿命预测理论体系，为多阶段随机退化设备的剩余寿命预测与健康管理提供有价值的理论与技术支撑。

基于状态监测信息准确预测其剩余寿命，并据此对设备实施健康管理，对于切实保障复杂设备的运行安全性、可靠性与经济性具有重要意义。针对现有研究主要集中在单一运行环境、单一工况下连续随机退化设备的建模及剩余寿命预测问题，本项目开展了基于多阶段随机退化过程建模的剩余寿命预测与健康管理理论与方法研究。项目执行期间，项目组按照预定计划开展了一系列相关研究，顺利完成了本项目的各项研究内容，主要包括：1）带随机突变的多阶段退化建模与剩余寿命预测方法；2）多阶段非线性随机退化建模与剩余寿命预测方法；3）考虑贮存退化下的预测维护与备件供给联合决策方法；最后，依托某型导弹控制系统关键设备实际数据以及其他公开数据集，开展应用验证。.在国家自然科学基金（No. 61903376）的支持下，项目组按照计划书中既定的研究内容开展研究，并完成了计划的研究目标，在国内外重要学术期刊和国际学术会议上发表论文13篇，其中SCI收录论文10篇、EI收录论文2篇（均标注本基金资助）；申请国家发明专利10余项、授权9项；协助培养研究生7名；获2022年陕西高等学校科学技术研究优秀成果一等奖1项（排名3），完成了项目的预期成果。本课题的研究有助于丰富随机退化设备的剩余寿命预测理论体系，已取得的成果可以为多阶段随机退化设备的剩余寿命预测与健康管理提供有价值的理论与技术支撑。

内容获取失败，请点击重试