Issues Concerning Material Constitutive Laws and Parameters in Springback Simulations

With the increased application of high strength steels in automotive body-in-white parts for weight reduction purposes, more emphasis is focused on springback as a major problem in stamping operations, in addition to panel breakage and wrinkling. Computer simulations using the finite element analysis (FEA) have been used to predict springback during early stages of die development processes to minimize potential springback related problems in production. However, the reliability of the springback simulation results relies directly on the accuracy of stress distributions from the forming simulation. Its complexity has brought many challenges not only to engineers and researchers in areas of FEA development and material modeling but also to FEA code end users. It is shown from this study that the springback simulation results vary with the yield criterion used in the forming simulation. The selection of different yield criteria may result in an entirely different springback mode for a given part. Springback simulation results are also very sensitive to the anisotropy values when an anisotropic yield criterion is used. For a panel undergoing bending followed by unbending or reversed deformation, kinematic hardening rule may result in different springback amounts when compared to isotropic hardening rule. Whenever possible, a set of complete stress and strain data should be used instead of using power law parameters such as n-value (strain hardening exponent) and K-value (material strength coefficient). It has also been demonstrated in this study that the strain rate sensitivity has a significant effect on springback prediction results.