融合新旧产品退化信息的可靠性建模研究  

Research on Reliability Modeling by Integrating Old and New Product Degradation Data Information

作　　者：吴怡 唐家银[1] 王劲博 刘新玲 WU Yi;TANG Jiayin;WANG Jinbo;LIU Xinling(School of Mathematics,Southwest Jiaotong University,Chengdu 611756,China)

机构地区：[1]西南交通大学数学学院,四川成都611756

出　　处：《机械与电子》2024年第4期9-14,共6页Machinery & Electronics

基　　金：西南交大首期来华留学全英语授课精品课程项目(LHJP[2023]10);教育部人文社会科学研究规划基金项目(20XJAZH009);西南交通大学教育教学研究与改革项目重点项目(20220320)。

摘　　要：针对2阶段退化产品截尾数据难以进行精确可靠性评估问题,利用贝叶斯融合新产品截尾数据和相似产品2阶段退化数据信息,并采用变系数分段回归模型和变点思想建立了变系数分段退化可靠性评估模型。首先,运用泰勒展开式和遗传算法得到相似产品回归模型的参数估计值并拟合分布。其次,通过贝叶斯方法融合新旧产品第1阶段退化数据信息,基于变点的连续性建立阶段关系式得到第2阶段新产品参数估计量期望和方差。最后,利用相似产品退化量的分布类型和矩估计方法得到新产品退化量的分布以及分段可靠度函数。算例分析结果验证了该模型的有效性和精准性。A reliability evaluation model was established for two stage degenerate data with variable coefficients.This model utilized Bayesian fusion to integrate new product Truncated data and two stage degenerate data from similar products.It also incorporated a variable coefficient piecewise regression model and the concept of change points.Firstly,the parameter estimates for similar product regression models were obtained by using Taylor expansion and a genetic algorithm,and then their distributions were fitted.Next,through integration with Bayesian methods,the expectation and variance of the new product parameter estimators in the second stage were obtained by combining the first stage degradation data of both new and old products.The model also took into account the continuity of the change points when establishing the stage relationship.By applying the distribution type and moment estimation method of similar product degradation,the distribution of the new product degradation and the piecewise reliability function were obtained.Finally,an example analysis was conducted to verify the effectiveness and accuracy of the model.

关 键 词：变系数分段回归模型 截尾退化数据 贝叶斯方法 变点连续性 分段退化可靠性评估模型 

分 类 号：TB114.3[理学—概率论与数理统计]