COMPARISON OF LOCAL AND GLOBAL APPROXIMATIONS FOR RELIABILITY ESTIMATION

An approach based on approximation concepts to estimate system reliabi lity is presented, Limit-state functions are approximated explicitly i n terms of the random variables of the system, These approximations ar e then used in conjunction with Monte Carlo simulation to estimate sys tem reliability, Local and global approximations are studied and compa red in terms of efficiency and accuracy for their potential use in rel iability-based optimization. Example problems are presented to illustr ate the ideas set forth.