Crashworthiness design of foam-filled bitubal structures with uncertainty

Abstract

Structural optimization has been widely used to improve the crashworthiness of foam-filled thin-walled structures. However, majority of the existing optimization studies to date have not considered uncertainties for simplication. Its associated risk is that a deterministic optimization might deteriorate its optimality and/or violate design constraints when being present in uncertain environment. In this study, a multiobjective robust design optimization (MORDO) method is adopted to explore the design of foam-filled bitubal structures. To reduce the computational burden of highly-non-linear crash analysis, adaptive Kriging models are employed in the optimization process. In this strategy, sequential sampling points are generated over the design space and Kriging models are refitted in an iterative fashion. Based on the Kriging models, the multiobjective particle swarm optimization (MOPSO) algorithm is employed to perform the optimization, integrated with Monte Carlo simulation and descriptive sampling technique. The results demonstrate that the proposed method is capable of improving the robustness of Pareto solutions within the prescribed minimum requirements of reliability. Moreover, the influence of varying the emphasis on mean and standard deviation components is also analyzed, which can provide decision-makers with insightful design information.