Improving further sensitivity analysis of possibilistic reliability of fuzzy structure

We at NWPU have continued for several years applying and improving further the method of possibilistic reliability sensitivity analysis proposed by us in Refs.3 through 6. Section 2 of the full paper derives two analytical PRS (possibilistic reliability sensitivity) expressions (as given in eqs. (13) and (14) in the full paper) for fuzzy structure for the special case of linear limit-state function with fuzzy variables possessing only Gaussian or only linear possibility distribution function. Section 3, guided by the above-mentioned analytical PRS expressions and using the linearization of non-linear limit-state function and equivalent transformation from non-Gaussian possibility distribution function to Gaussian one, establishes an approximately analytical method for calculating the PRS value for the general case. Subsection 4.1 gives one new numerical method for calculating PRS value and subsection 4.2 gives another new one. Section 5 presents six numerical examples of possibilistic reliability sensitivity analysis. The PRS analysis results, shown in Tables 1 through 6, indicate preliminarily that the two new numerical methods can effectively calculate the PRS value of the fuzzy structure and that the calculation results are close to those of the analytical solutions (Examples 1 through 5 and Tables 1 through 5) and those of the approximately analytical method (Example 6 and Tables 6.1 and 6.2).