摘要
本文研究了并联结构体系可靠度分析中以失效概率估计值方差最小为条件的重要抽样问题。由于直接对抽样样本平均值和标准差进行优选比较困难,因而本文首先建立了三个定理,据此并通过数例分析给出正态抽样样本平均值和标准差的建议取值范围。分析和实际模拟表明,在以正态分布进行抽样的前提下,合理选取抽样样本平均值和标准差,可使抽样效率进一步提高。
The importance sampling method which minimizes the variance of structural systems failure probability estimator is studied.Because of the difficulty to obtain the optimum mean value and standard deviation of sampling variables, three theorems are at first founded, then the range of mean value and standard deviation of sampling variables are proposed based on these three theorems and results of some examples.Theorical analysis and practical simulation indicate that if the samples are of normal distribution, the modelling efficiency can be increased further by using sensible parameters.
出处
《工程力学》
EI
CSCD
北大核心
1997年第3期29-35,共7页
Engineering Mechanics
基金
国家基础性研究重大项目
关键词
结构体系
可靠度
蒙特卡洛模拟
最小方差
structural systems, reliability, Monte-Carlo simulation, importance sampling, minimum variance