摘要
针对非理想贮存条件下备件退化甚至失效,导致运行设备与备件组成的贮备系统实际运行寿命难预测的问题,提出了一种考虑备件性能退化的贮备系统剩余寿命预测方法。基于维纳过程分别建立了系统在运行和贮存条件下的随机退化模型,引入了随机效用刻画系统退化过程的个体差异性,推导出备件的时变退化状态分布及失效概率;结合运行条件下与贮存条件下的退化模型,推导出了首达时间意义下贮备系统剩余寿命分布的迭代求解方法;基于极大似然估计和贝叶斯理论,提出了模型参数的离线辨识和在线更新策略。通过数值仿真数据和某型号陀螺仪的实际数据进行实验验证,实验结果表明:所提方法的剩余寿命期望理论值与数值仿真结果高度吻合,且理论值与近似值的平均相对误差为0.030 3 h,验证了所提方法的理论正确性;同时,所提方法能够有效表征备件退化对于剩余寿命预测的影响,剩余寿命与真实剩余寿命的平均均方根误差为37.645 3 h,比不考虑随机效用的方法降低了1.004 h,预测精度显著提高。该方法为贮备系统剩余寿命预测提供了有效途径。
To address the challenge of predicting the actual operational life of a standby system,comprising operating equipment and spare parts,due to the degradation or failure of spare parts under non-ideal storage conditions,a method for estimating the residual useful life of standby systems while considering spare part degradation is proposed.Stochastic degradation models for the system under both operational and storage conditions are developed using Wiener processes.Individual differences in system degradation are accounted for through stochastic utility,enabling the derivation of the time-varying degradation state distribution of spare parts and failure probabilities.By integrating degradation models for operational and storage conditions,an iterative approach for determining the distribution of the system’s residual useful life concerning the first reach time is established.An offline parameter identification and online updating strategy,employing maximum likelihood estimation and Bayesian theory,are introduced.Experimental validation is performed utilizing numerical simulations and empirical data from a specific model gyroscope.The experimental results reveal that the expected theoretical values of residual useful life generated by the proposed method closely align with numerical simulations,with an average relative error of 0.0303 h between theoretical and approximate values,validating the method’s theoretical accuracy.Moreover,the impact of spare part degradation on residual useful life predictions is effectively assessed by the method,with an average root mean square error of 37.6453 h between predicted and actual residual useful life achieved,representing a reduction of 1.004 h compared to approaches neglecting stochastic utility and significantly enhancing prediction accuracy.A robust means of forecasting the residual useful life of standby systems is offered by this method.
作者
张迦陵
张建勋
杜党波
张正新
司小胜
ZHANG Jialing;ZHANG Jianxun;DU Dangbo;ZHANG Zhengxin;SI Xiaosheng(Zhijian Laboratory,Rocket Force University of Engineering,Xi’an 710025,China)
出处
《西安交通大学学报》
EI
CAS
CSCD
北大核心
2024年第12期175-185,共11页
Journal of Xi'an Jiaotong University
基金
国家自然科学基金资助项目(62373368,62203462)。
关键词
剩余寿命
退化过程
首达时间
贮备系统
备件退化
remaining useful life
degradation process
first passage time
standby system
spare parts degradation
