基于ALT和MCMC方法的产品外场可靠性评估方法  

Product Field Reliability Assessment Based on ALT and MCMC Method

作　　者：张国龙[1] 蔡金燕[1] 梁玉英[1] 

机构地区：[1]军械工程学院电子与光学工程系,石家庄050003

出　　处：《航天控制》2015年第2期93-97,共5页Aerospace Control

基　　金：国家自然科学基金(61271153)

摘　　要：针对目前基于加速试验产品可靠性分析方法忽略外场环境的差异性及未考虑外场数据等问题,从数据融合的角度,考虑试验与外场环境的差异性,综合加速寿命试验的试验数据和现场数据2种信息,建立更为准确的模型,并利用马尔科夫链蒙特卡洛(MCMC)进行贝叶斯分析,解决模型复杂难于计算的问题,通过Gibbs抽样获得模型后验参数估计,从而对产品在外场工作状况下的寿命及可靠性进行评估,最后通过实例对比分析,说明综合2种数据来源不仅扩大了信息量,同时评估的精度更高,结果更加可信。According to improve the reliability assessment accuracy, in view of information fusion, a more accurate model is established by integrating with failure information from both the field data and accelerated life testing data, and Marcov Chain Monte Carlo method is used for Bayesian analysis and posterior param- eter estimation by Gibbs sampling, which can effectively reduce the computational complexity of the model. Then the life and reliability of the product with field condition can be obtained. Finally, a demonstrated ex- ample is simulated and comparatively analyzed, and the result shows that reliability assessment based on the integration of ALT and field data is more precise and credible, and also has the information expanded.

关 键 词：加速寿命试验 外场可靠性 马尔科夫链蒙特卡洛 环境差异因子 

分 类 号：TB114.3[理学—概率论与数理统计]