摘要
复杂系统可靠性增长试验是一个动态的过程,成本所限不允许进行大规模的试验,而传统方法在子样容量不足的情况下误差较大,无法精确地反应系统可靠性增长的特点,因此在小子样条件下对复杂系统可靠性增长进行评估是目前亟待解决的问题。基于AMSAA模型下复杂系统可靠性增长试验的特点,结合少量现场试验数据,提出了可靠性增长评定的动态建模方法,运用Bayes线性回归方法对系统的可靠性参数进行预测,并对可靠性增长试验的结果进行评定,并给出了各阶段可靠性增长试验中可靠性参数的Bayes估计。由仿真实例证明,与传统方法相比,文中的动态建模方法更加准确,能够有效的解决小子样条件下的可靠性增长问题。
The reliability growth test evaluation of complex system is a dynamic process, and its cost is too huge to be tested in large sample circumstance. The error of evaluation in classic method is too big to reflect the characteristics of reliability growth, so, how to complete the evaluation in small sample circumstance is a problem demanding prompt solution. Based on the characteristics of complex system reliability growth test and combined with a few field test data, Bayesian linear regression method is used to forecast the reliability parameters and evaluate the reliability parameters of systems under the AMSAA model in this paper. With the characteristics of invalidation probability and reliability growth test data, a dynamic parameter iterative evaluation model is built in this article and the reliability evaluation of every stage is given out by Bayesian formula. Compared with the classical method, the dynamic linear model is more correct and more valid to solve the evaluation in small sample circumstance.
出处
《计算机仿真》
CSCD
2007年第11期312-315,共4页
Computer Simulation
关键词
复杂系统
可靠性增长
动态线性模型
Complex system
Reliability growth
Dynamic linear model
作者简介
吴祺(1981-),男(汉族),山东青岛人,国防科学技术大学机电工程与自动化学院硕士研究生,主要研究方向:武器装备试验鉴定,可靠性研究;
闫志强(1981-),男(汉族),河南焦作人,国防科学技术大学机电工程与自动化学院博士研究生,主要研究方向:可靠性研究,故障诊断;
谢红卫(1965-),男(汉族),湖北洪湖人,国防科学技术大学机电工程与自动化学院博士生导师,教授,主要研究方向:系统工程,可靠性工程,生物信息工程。