Finite element simulation of the manufacturing process chain of a sheet metal assembly

An increasing number of components in automotive structures are today made from advanced high strength steel (AHSS). Since AHSS demonstrates more severe springback behaviour than ordinary mild steels, it requires more efforts to meet the design specification of the stamped parts. Consequently, the physical fine tuning of the die design and the stamping process can be time consuming. The trial-and-error development process may be shortened by replacing most of the physical try-outs with finite element (FE) simulations of the forming process, including the springback behaviour. Still it can be hard to identify when a stamped part will lead to an acceptable assembly with respect to the geometry and the residual stress state. In part since the assembling process itself will distort the components. To resolve this matter it is here proposed to extend the FE-simulation of the stamping process, to also include the first level sub-assembly stage. In this study a methodology of sequentially simulating each step in the manufacturing process of an assembly is proposed. Each step of the proposed methodology is described, and a validation of the prediction capabilities is performed by comparing with a physically manufactured assembly. The assembly is composed of three sheet metal components made from DP600 steel which are joined by spot welding. The components are designed to exhibit severe springback behaviour in order to put both the forming and subsequent assembling simulations to the test. The work presented here demonstrates that by using virtual prototyping it is possible to predict the final shape of an assembled structure.

► A methodology to predict the properties of a sheet metal assembly is presented.
► Each step of the manufacturing process is simulated sequentially.
► The evolution of deformations through the manufacturing process chain can be followed.
► For validation purpose an assembly was manufactured, both physically and virtually.