| Meta-modeling has been widely used in place of complex numerically intensive simulations to perform design reliability assessment and optimization. Due to cost and time constraints, especially when uncertainty is considered, most complex simulations can only afford a limited number of runs with a relatively large number of factors. The goal of this dissertation is to advance the use of Metamodeling techniques in engineering design by developing efficient and robust methods for constructing metamodels and further exploiting metamodels to gain design knowledge and to facilitate uncertainty analysis. Research developments have been made in the following three main areas.; In the area of optimal design of computer experiments, the effect of DOE optimality criteria on the accuracy of the Kriging meta-model is investigated under scarce sampling plans. Based on a comprehensive statistical comparison for real world case studies, it is found that the Maxmin and entropy criteria is more efficient than other existing optimization criteria.; In the area of sensitivity analysis, an efficient method using moment matching is developed for estimating the impact of each design parameter on response variance. Numerical examples showed that the proposed method provided accurate results and requires less runs than existing techniques.; Finally, a new search method for the Most Probable Point (MPP) which is a good metric for measuring product reliability is proposed. The new method is based on a simulation within a limited radial sampling region determined by an initial uniform DOE that guarantees a good coverage of design space. Numerical nonlinear examples are used to compare and verify the validity of the new method.; Through example problems from literature and industrial applications, it is illustrated that the research developments in this dissertation can be integrated to provide efficient and reliable techniques for advancing the application of Metamodeling techniques in real world engineering design problems. |
