Search Results

Automotive body frame structures play crucial roles in the achievement of dynamic and crashworthiness performance, and their cross-sections must fulfill various required characteristics. However, cross-sectional shape design is complex and relies on designers' intuitions, since automotive body frames are fabricated from many sheet metal parts having complex interdependencies, a problematic situation for optimization via software. This paper proposes a cross-sectional shape generation method satisfying a range of required cross-sectional properties for early design stages. Shape generation problems are formulated as multiobjectsve optimization problems, solved using genetic algorithms and a variety of cross-sections can be obtained as Pareto optimal solution sets.

This paper proposes a new hierarchical optimization technique for the vehicle body structure, by combining topology optimization and shape optimization based on the traction method. With the proposed approach, topology optimization is first performed on the overall allowable design domain in 3D. The surface is extracted from the optimization result and converted to a thin shell structure. Shape optimization based on the traction method is then applied to obtain an overall optimal body shape. In the shape optimization process, iterative calculations are performed in the course of consolidating parts by deleting those whose contribution is small. The result obtained by applying this method to the front frame structure of a vehicle is explained. The resultant optimal shape has stiffness greater than or equal to the original structure and is 35% lighter. This confirms the validity of the proposed technique. It was found, however, that some issues remain to be addressed.