摘要
竞争失效场合加速试验技术是加速试验由简单结构产品向复杂结构产品推广应用的基础,而如何设计试验方案使统计结果最准确、代价最小,是其中的主要研究内容之一。针对传统解析优化方法推导过程比较复杂的问题,提出了一种基于Monte Carlo仿真的竞争失效场合加速试验优化设计方法。在备选方案较多的情况下,通过引入曲面拟合进行间接优化,这样可以大量减少进行仿真的试验方案个数,进而在保证较高精度的同时,使得计算量大大减小。通过两个算例分别演示仿真基加速试验方案设计的直接优化和间接优化方法,表明方法易于流程化、适合工程应用。最后,敏感性分析结果表明该方法具有一定的鲁棒性。
The technology of accelerated testing with competing failure is the foundation of application extending from products with simple structure to products with complex structure.Designing optimal test plans,which obtains better result with lower cost,is one of the main research points in the technology.Traditional analytical optimization has some shortcomings,such as complexity of deduce process.To overcome these shortcomings,this study presents a new method of Monte Carlo simulation based optimal designs for accelerated testing with competing failure.When the candidate plans are too many,curve fitting can be introduced to decrease the amount of calculation by reducing the number of test plans for simulation.Two cases of simulation based optimal designs are demonstrated by direct and indirect optimization respectively,which shows that the proposed method is suitable for application.Results of sensitivity analysis show that this method is robust.
出处
《国防科技大学学报》
EI
CAS
CSCD
北大核心
2011年第2期130-135,共6页
Journal of National University of Defense Technology
基金
国家部委资助项目(203020102)
