A truncated reliability analysis method with the fuzzy boundary

This paper proposes the fuzzy truncated probabilistic distribution method to study fuzzy information of design variables. Instead of the certain boundary of design variables, in practical measurements, the fuzzy boundary tends to be employed to represent the variables range, in which some measurement errors can be considered to perfect the traditional reliability model. Furthermore, the fuzzy reliability model is established with generalized resistance and generalized stress successfully, which can be able to meet the requirement of engineering ap-plications. The calculation process, however, tends to fall into a dead end because it is difficult to clearly describe the integration domain by direct integration method. Thus, a corresponding improved algorithm is also proposed to achieve it. As a reliable numerical integration method, the result of reliability analysis can be obtained by Gauss-Legendre quadrature, for the value of the integral operation is available in a certain lambda-cut set level. Eventually, three numerical examples are investigated to demonstrate the feasibility and effectiveness of the present method, where it is proven that the algorithm improves the calculation efficiency compared with the traditional Monte Carlo Simulation (MCS) method.