Springback Simulation of the Numisheet 2005 Benchmark II Using DP600

This study uses a design of experiments (DOE) methodology to investigate the sensitivity of springback prediction to various numerical parameter for the Numisheet 2005 cross member (Benchmark II). The parameters investigated are; through thickness integration points (21 Gaussian integration points through the thickness vs. 7), using a static implicit finish in the forming simulation, element size (number of adaptive levels), and coulomb friction. The average effect of these parameters on the resulting springback was then quantified. Overall, it was found that element size and friction have the greatest effect on the predicted springback and that little is gained by using an implicit finish and 21 integration points through the thickness. Possible reasons for this are discussed and it is then stressed that, these results cannot be generally applied to all situations. In other words, the results point to the possibility that some parts (such as this one) are not sensitive to increasing integration points (from 7 to 21) or an implicit finish. It was also found that despite all combinations of numerical parameters, the wall curl on one side of section I of the Benchmark could not be reproduced