加速度计步降应力加速寿命试验优化设计  

Accelerated life testing optimal design of step-down-stress accelerometer

作　　者：罗赓 穆希辉 牛跃听 姚安[4] 王永南[2] 

机构地区：[1]中国华阴兵器试验中心,陕西华阴714200 [2]军械工程学院三系,河北石家庄050003 [3]军械技术研究所,河北石家庄050003 [4]军械工程学院五系,河北石家庄050003

出　　处：《机械设计》2016年第4期78-83,共6页Journal of Machine Design

基　　金：国家自然科学基金资助项目(61471385)

摘　　要：针对解析方法难以得到或者不能得到步降应力加速寿命试验最优方案的难题,提出了一种基于Monte-Carlo仿真的步降应力加速寿命试验(SDSALT)优化方法。该方法通过大量的重复模拟试验生成试验失效数据,寻找最佳应力水平和试样分配比例,以正常使用应力下的P阶分位寿命均方误差估计(MSE)和各个应力水平下的特征寿命之和最小为目标,采用极大似然估计(MLE)对失效数据进行统计分析,建立基于仿真的步降应力加速寿命试验优化设计模型。给出的算例仿真结果说明该方法的可行性与正确性,易于流程化,适合工程应用。Aiming at the problem of an optimal scheme for step-down-stress accelerated life test（SDSALT） cannot be obtained through analytical method, an algorithm of optimal design based on Monte-Carlo simulation was presented. The algorithm was used to find the optimized stress level and sample proportion using many times simulation test. The mean square error estimation of 100 pth percentile of the p steptile product lifetime distribution at normal condition and the sum of sample life under different stress level was minimum as the objective. The maximum Likelihood Estimation（MLE） theory was used to take statistical analysis and established the optimal design model of SDSALT based on simulation. The result of an example given proved that this algorithm was feasible and precise and met application requirement.

关 键 词：蒙特卡洛仿真 加速寿命试验 步降应力 寿命均方误差估计 优化设计 

分 类 号：TP114.3[自动化与计算机技术—控制理论与控制工程]