摘要
通过对可靠性灵敏度的分析和推导,利用灵敏度标准差值构造了一种新的可靠性灵敏度因子。与传统灵敏度因子不同的是,新因子不仅能反映各随机变量对结构失效概率影响的重要性程度,因子的数值大小还能表征将单个变量作为确定性变量处理或将多个变量同时作为确定性量处理时所引起的可靠指标误差大小。据此结论,采用新灵敏度因子进行灵敏度分析时,只需要进行一次分析就能识别出各随机变量对结构失效整体影响的有益信息,可以有效地提高计算效率,降低分析难度。利用数值算例验证说明了所提可靠性灵敏度因子指标的有效性。
A new reliability sensitivity factor is constructed by using the standard deviation of reliability sensitivities, based on the analysis and derivation of reliability sensitivities. Different from the traditional sensitivity factor, the new one can reflect the importance difference of all stochastic variables to structural failure probability, and also can represent the error caused by a variable or multiple variables considered as a constant. According to the conclusion, all helpful information from the effect of all stochastic variables on structural failure probability can be recognized by performing a reliability sensitivity analysis. This method can thusly improve the calculation efficiency and reduce the analysis difficulty greatly. The paper illustrates and verifies the conclusion by numerical examples in the final.
出处
《工程力学》
EI
CSCD
北大核心
2013年第6期16-21,29,共7页
Engineering Mechanics
基金
国家自然科学基金青年科学基金项目(11002033)
高等学校博士学科点专项科研基金项目(20100041120013)
关键词
可靠性
灵敏度因子
可靠指标
分布参数
二阶矩法
reliability
sensitivity factor
reliability index
distributed parameter
second moment method