Efficient safety index calculation for structural reliability analysis

This paper proposes an efficient structural reliability analysis algorithm with an adaptive nonlinear approximation of performance function, in which the nonlinearity of the function is controlled by the feedback information from the previous iteration. The proposed analysis procedure includes (1) identifying the most probable point of a limit state, (2) establishing an approximate nonlinear performance function by using intervening variables and (3) applying the HL-RF or the modified HL-RF method. The method is suitable not only for problems that involve explicit performance functions, but also for problems that need computer-intensive structural analysis. Several examples are presented in this paper to demonstrate the performance of the analysis method using large coefficients of variation and nonnormal distribution of random variables.