Abstract: The distribution parameters of a probabilistic model are usually obtained from statistic inference of random experiment results, and they cannot generally be forecasted with accuracy. It is therefore reasonable and feasible to describe them as fuzzy variables. By combining fuzzy mathematical theory and reliability analysis of structural systems, this article makes an analysis of the fuzzy reliability of structural systems with fuzzy variables as distribution parameters of a probabilistic model. The fuzzy correlation coefficient of any two failure modes and the solution formula of secondorder fuzzy failure probability are deduced in detail, and λ cut numbers of the fuzzy failure probability of the structural system is carried out when the failure modes are either correlated or uncorrelated. Fuzzy failure probability of the structural system can finally be estimated. Fuzzy reliability of a stiffening plate system is analyzed as an application of the proposed method. Finally, a numerical example is considered to illustrate the practicability of the method.

Key words: structural system, distribution parameters, fuzzy variable, fuzzy failure probability, reliability, stiffening plate