基于Copula函数耦合性建模的二元加速退化数据统计分析方法

Statistical Analysis Method of Bivariate Degradation Data Based on Dependency Modeling Via Copula Function

以某型继电器为研究对象,提出了一种二元加速退化数据建模方法:利用Wiener过程建立性能参数的退化模型,然后结合阿伦尼斯方程建立模型参数的加速退化模型,采用Copula函数建立二元加速退化过程之间的耦合性模型;为了一体化估计二元加速退化模型中的多个参数,设计了一种基于Bayesian马尔可夫链蒙特卡罗的参数估计方法;利用继电器二元加速退化数据统计分析实例验证了所提建模方法与参数估计方法的可行性;研究工作为解决类似产品的可靠性评估问题提供了有益借鉴,为完善二元加速退化数据统计分析理论做出了一定贡献。

A modeling method of bivariate accelerated degradation data was proposed. A Wiener process was used to establish degradation models for each degradation process, and then the Arrhenius equation was applied to establish accelerated degradation models for model parameters, finally Copulas functions were adopted to construct the dependency model between bivariate accelerated the degradation processes. A parameter estimation method based on Bayesian Markov Chain Monte Carlo was designed to estimate the muhiple parameters of bivariate accelerated degradation model. A case application of a certain type of relay was provided to validate the feasibility of the proposed methods. The research work may help solve other reliability assessment problems for similar products, and makes some contributions to develop the theory of statistically analyzing bivariate acceleration degradation data.