摘要
针对贵重产品可靠性评定中只有极少几个试验数据的情况,提出一种极小子样可靠性评定方法。该方法能将仿真(或数字化设计)结果与试验数据有机结合,实现多个不同状态下(多个母体)试验数据的整体推断,从而建立各状态下均值和方差的估计量,给出母体百分位值和百分率的置信限和置信区间估计。与传统方法相比,该方法具有更高的精度,解决了一种状态下只有一个试验数据时产品可靠性评定的难题。
A method for reliability assessment with a few experimental data of valuable products is presented. By combining the results of simulation (or digital design) and the experimental data, it can totally infer the experimental data in multi-state (for multi-population). And then in each state it can estimate not only the mean and the variance, but also the confidence limits and intervals for population percentiles and percentages. Compared with traditional methods, this presented method has higher precision and solves the problem of reliability assessment with only one experimental datum in a state.
出处
《机械强度》
EI
CAS
CSCD
北大核心
2005年第3期335-338,共4页
Journal of Mechanical Strength
基金
国防科技预研项目(41801060402)资助。~~
关键词
小子样
整体估计
可靠性
仿真
数字化设计
Small sample
Integral estimate
Reliability
Simulation
Digital design
