摘要
为探讨输入变量在随机不确定性环境下对产品的分位点(产品满足一定概率要求的输出性能的上限值)的影响,定义了基于分位点的输入变量的全局灵敏度分析(GSA)指标。该指标能够在给定的概率要求下全面衡量输入变量在其分布域中变化时对输出性能分位点的平均影响程度。揭示了该指标与已有的基于分布函数的全局灵敏度指标和基于失效概率的全局灵敏度指标的内在联系,并利用维度缩减方法和基于分数阶矩的极大熵算法以及Nataf变换来高效求解所提指标。通过数值和工程算例说明了基于分位点的全局灵敏度指标的物理意义,并验证了求解方法的精度和效率。
In order to investigate the effect of uncertainties of input variables on the quantile fractile the upper limit of the output performance of the products which satisfy a given probabilistic request,a global sensitivity analysis （GSA） index of the input variables based on the quantile fractile was defined.Under a given specific probability value,this sensitivity index can be used to completely assess the average effect of the input variables on the quantile fractile of the output performance when the input variables vary in their distribution ranges.Furthermore,the inherent relationship between the defined index and the existing distribution function （DF） based GSA index and the existing failure probability based GSA index was derived,and a method based on the concept of dimension reduction,a maximum entropy algorithm using fractional moments and the Nataf transformation method were used to efficiently calculate the proposed index.At last,one numerical example and two engineering examples were introduced to show the significant meaning of the proposed GSA index and demonstrate the precision and the efficiency of the proposed computational method simultaneously.
出处
《高技术通讯》
CAS
CSCD
北大核心
2014年第8期858-865,共8页
Chinese High Technology Letters
基金
国家自然科学基金(51175425)资助项目
关键词
灵敏度分析
分位点
分数阶矩
极大熵
维度缩减
Nataf变换
sensitivity analysis
quantile fractile
fractional moment
maximum entropy
dimension reduction
Nataf transformation
