随机冲击影响的非线性退化设备剩余寿命预测  

Remaining useful life prediction method for degradation equipment with random shocks

作　　者：白灿 胡昌华 司小胜 李洪鹏 张正新 裴洪 BAI Can;HU Changhua;SI Xiaosheng;LI Hongpeng;ZHANG Zhengxin;PEI Hong(Test Teaching and Research Section of Rocket Military Engineering University,Xi＇an 710025,China;Beijing Institute of Remote Sensing Equipment,Beijing 100854,China)

机构地区：[1]火箭军工程大学测试教研室,陕西西安710025 [2]北京遥感设备研究所,北京100854

出　　处：《系统工程与电子技术》2018年第12期2729-2735,共7页Systems Engineering and Electronics

基　　金：国家自然科学基金(61573365;60736026;61174030;61104223);国家杰出青年基金(61025014)资助课题

摘　　要：退化设备的剩余寿命(remaining useful life,RUL)预测是当前可靠性领域研究的一个热点问题。基于Wiener过程,提出一种考虑随机冲击影响的非线性退化设备RUL预测方法。首先,设备连续退化过程用一个非线性Wiener过程描述,而冲击导致退化水平突变的影响由一个复合泊松过程刻画;其次,基于所建立的退化模型和首达时间概念,推导出剩余寿命概率密度函数及其近似解析解,极大地减少了数值计算时间,并提出一种基于期望最大化算法的模型参数估计方法。数值仿真和航天锂电池实例验证表明,所提方法提高了RUL预测的准确性,对于解决存在随机冲击影响的设备RUL预测问题具有一定的理论指导意义。The remaining useful life（RUL）prediction of degradation equipment is a hot topic in the field of reliability research.Based on the Wiener process,a RUL prediction method considering random shocks is proposed here.Firstly,the continuous degradation process of the equipment is described by a nonlinear Wiener process,while the random shocks leading to the mutation of the degradation level is characterized by a compound Poisson process.Secondly,based on the established stochastic degradation model and the concept of the first hitting time,the approximate analytic probability density function of RUL are deduced,which greatly reduces the numerical computation time,and a model parameter estimation method based on the expected maximization algorithm is proposed.Numerical simulation and the aerospace lithium battery example indicate that,the proposed method improves the accuracy of RUL prediction,and it has certain theoretical guiding significance for solving the RUL prediction problem of equipment with random shocks.

关 键 词：随机冲击 非线性退化 寿命预测 期望最大化算法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]