Advanced high strength steel springback optimization by projection-based heuristic global search algorithm

The forming process of advanced high strength steel (AHSS) often suffers severe springback. Due to the nonlinear characteristics of springback, the expensive computational cost thereby is a major obstacle for finite element analysis-based springback optimization. Therefore, this paper suggests a metamodel-assisted optimization named as projection-based heuristic global search method (P-HGS) to minimize springback. The efficiency of the suggested method is remarkably improved. In this method, a general set of quantitative model assessment and analysis tool, termed high dimension model representation (HDMR) is employed to construct metamodel. Compared with conventional metamodels, HDMR can generate a more accurate metamodel with a smaller size of samples. In order to further enhance the efficiency, a heuristic sampling method-mode pursuing sampling (MPS) is adopted to generate samples. The paper also proposes a projection strategy to integrate HDMR and MPS seamlessly. The performance of the proposed approach is verified by several nonlinear test functions. Drawbead and blank holder force (BHF) design in sheet forming process are successfully optimized for springback by the P-HGS method.

► The Kriging-based Cut-HDMR technique is used to improve the performance of metamodel-assisted optimization.
► Projection strategy integrates Kriging-HDMR to the sampling method seamlessly.
► Two typical springback problems are investigated by the proposed P-HGS method.