基于Copula函数的失效概率灵敏度分解方法

For the sensitivity analysis of failure probability with dependent input variables, by introducing the Copula function to describe the joint probability distribution of input variables, the sensitivity of failure probability can be decomposed into independent sensitivity and dependent sensitivity. Then for the computation of the sensitivity of failure probability, the derivative of failure probability can be transformed into the kernel function of marginal distributions and Copula functions. While for the common probability distributions and Copula functions, their kernel functions can be derived analytically, which is directly used to the computation of sensitivity of failure probability. The numerical examples and engineering practice are employed to validate the rationality of the proposed method.