Simulation of springback with the draw/bend test

This paper summarizes a length analysis of springback of the draw/bend test, conducted using three sheet materials, several friction coefficients, die radii, and draw-in restraining forces. Early results showed large discrepancies between experiments and simulations for some conditions. The simulations have been optimized since that time and their sensitivity to a variety of numerical parameters and, more recently, to the choice of finite element and material model, have been examined. The finite element analysis (FEA) of springback is shown to be very sensitive to numerical parameters, including the number of through-thickness integration points, the angle of contact per shell element, and the tolerances for equilibrium and contact. With the help of numerical sensitivity studies, guidelines are provided for choosing these values effectively. Good agreement between experimental and simulated (3D FEM modeling) springback has now been obtained for a range of process parameters. From this further analysis, it is concluded that the presence of 3D bending (anticlastic curvature) effects dominate the discrepancies, with smaller errors caused by the material model.