基于中值矩立方正态变换的结构可靠度分析OA

Reliability assessment is critical for structural design and maintenance.The traditional higher-order moment method has been widely used due to its simplicity and ease of implementation.However,when characterizing the left-tail probability features of a performance function using higher-order central moments,the method is often affected by deviations in the right tail,which may lead to inaccurate reliability indices.To address this issue,this study introduces median moments to capture the one-sided tail characteristics of the distribution,and combines them with a cubic normal transformation to establish anew reliability index formulation.A Markov Chain Monte Carlo(MCMC)approach is employed to estimate the median moments of the performance function.Numerical examples demonstrate that the proposed method provides high accuracy in solving engineering problems involving high-dimensional random variables,strongly nonlinear performance functions,and low failure probabilities.