Nondeterministic optimization of tapered sandwich column for crashworthiness

Abstract

Foam-filled thin-walled structures signify a class of a promising energy absorber for improving the crashworthiness and safety of vehicles. Although the conventional deterministic optimization has been extensively applied to crashworthiness design of foam-filled thin-walled structures, the optimal solution could become infeasible when uncertainties of design variables and noise factors present in real world. To address this issue, a reliability based design optimization (RBDO) is adopted to consider the uncertainties of design variables and noise factors in crashworthiness optimization for the foam-filled bitubal tapered structure in this paper. Moreover, to comprehensively investigate the differences between deterministic and reliability based design optimization, single objective and multiple objective RBDO are established by integrating Kriging approximation with Monte Carlo Simulation (MCS). Since the optimal results of deterministic design usually converge at the constraint boundary, the solutions of RBDO often need to compromise some objective performance to satisfy the predetermined reliability levels. Furthermore, a comparative study on different Pareto fronts yielded from the deterministic optimization and RBDO under different reliability levels is conducted here. Besides, a grey relational analysis is carried out to determine the most satisfactory solution from the Pareto-set. The results demonstrate that the optimized foam-filled bitubal tapered columns are capable to considerably improve capacity of energy absorption with an increased reliability, potentially being a structural configuration for energy absorber.