A time-space Kriging-based sequential metamodeling approach for multi-objective crashworthiness optimization

A time-space Kriging-based sequential metamodeling approach is proposed for multi objective crashworthiness optimization (MOCO) in this paper. By defining the novel time space design criteria, the constructed metamodels for the optimization objectives include the characteristic mechanical responses with respect to both the structural space domain and crash time domain, compared to standard metrics with the extremum of the time history of the entire structure. The adaptive addition of new samples is performed to gradually improve the approximation accuracy during the optimization with the guidance of an adaptive weighted sum method. The effectiveness of the proposed method is demonstrated by investigating a multi-cell thin-walled crashworthiness design problem. Finally, its effectiveness in practical engineering is validated by the crashworthiness design for a vehicle under full-overlap frontal crash loadcase. (C) 2018 Elsevier Inc. All rights reserved.