Dynamic covariance reliability analysis and residual life prediction model under the correlation between initial state and degradation rate
WEI Xin
TANG Jiayin
SONG Qingbiao
LI Wenxue
Abstract：[Objective]In the field of engineering technology,the degradation of systems is affected by various factors such as user operations,manufacturing processes,and working environments.The correlation between the initial state and the degradation rate(CISDR)and dynamic covariates are key factors to be considered in degradation models;however,there is little literature that takes both types of randomness into account simultaneously.To address the above issue,a reliability model based on the Gamma degradation process,which considers dynamic covariates and the correlation between the initial degradation state of the system and the degradation rate,was established for research and analysis.[Methods]Firstly,a residual life prediction model was established.Based on the statistical analysis of condition monitoring data during system operation,the residual life distribution of the product was further inferred.Secondly,in view of the differences in operation and operating environment during the system degradation cycle,the Ornstein-Uhlenbeck(OU)process was used to characterize the changes in dynamic covariates,and a Gamma degradation model considering dynamic covariates was constructed.Thirdly,the relation between dynamic covariates and degradation rate was established through an exponential effect model.Finally,a bivariate normal distribution was employed to build a correlation model between the initial degradation state of the system and the degradation rate,and the system reliability function and the probability density of residual life were derived.[Results]The results show that both the simulation examples and application cases have verified that the established model can significantly improve the accuracy of residual life prediction.Meanwhile,the residual life prediction that takes two types of random effects into account is more objective.
Keywords：Random initial stateDynamic covariateOU processResidual useful life predictionGamma process
Publication Date：2026-03-15
Online Publishing Date：2026-09-12(First online date of this platform, not the publication date of the document)
Pages：11( 10-20 )