Abstract: Targeting the problem of limit state equations being often implicit and demanding a great amount of computation, an analysis method was proposed for structural reliability sensitivity based on the bisection method of sampling. Combining with the most probable point method of structure reliability analysis and the bisection method of sampling, the proposed method first linearizes the state equations near the initial point based on Taylor expansion method and determines the optimal step-length based on gradient information; then, it locates the new sampling point by means of bisect sampling step-length of reliability index β; finally, by several iterations, it finds out the most probable point (MPP) that meet the accuracy requirement of the structures and system to calculate reliability and reliability sensitivity of influencing parameters according to the information of the sampled path. Examples of the computation and engineering were given. A comparison with existing methods for solving implicit structure reliability problems indicated that the proposed method possesses the advantages of reliable convergence and less sampling, making it especially suitable for performing reliability analyses of large-scale complicated implicit structures and systems.

Key words: Taylor expansion method, reliability index, most probable point, reliability sensitivity analysis