A New Hybrid Simulation Approach to Structural Reliability Analysis Using Uniform Design, ANN Meta-model, Genetic Algorithms and FORM
A new hybrid simulation method for structural reliability analysis is proposed which combines uniform design (UD) technique, artificial neural network (ANN) based meta-model, genetic algorithm (GA) and first order reliability method (FORM).The uniform design instead of classical central composite design (CCD) or orthogonal array design (OAD) is applied to choose experiment points in space of basic random variables aimed to minimize the number of simulation and to fill the space more uniformly.A BP-ANN based meta-model is used as a smart response surface surrogate to the original implicit limit state function in the global random variable space, with the UD experimental points as input training data sets of the ANN.Due to the highly nonlinear nature of ANN-based smart response surface, the Genetic algorithm (GA) incorporating FORM is employed to search for the global design point or most probable point (MPP) of failure to avoid fall into the local optimal solutions.To implement deterministic finite element analysis in the evaluation of the limit state function and finite element response sensitivity, the proposed approach is programmed in MATLAB by calling and integrating the commercial finite element analysis program ANSYS.Three numerical examples are provided to demonstrate the accuracy, efficiency and applicability of the proposed method by contrasting the new approach with the classical computational reliability methods such as Monte Carlo simulation (MCS), first order reliability method (FORM), and response surface method (RSM).
structural reliability uniform design artificial neural network genetic algorithms first order reliability method response surface method Monte Carlo simulation (MCS)
Dagang Lu Boqiang Li
School of Civil Engineering,Harbin Institute of Technology,150090,Harbin,China
国际会议
黄山
英文
1-10
2012-06-18（万方平台首次上网日期，不代表论文的发表时间）