An Investigation of Drawbead Models for Springback Prediction

Accurate prediction of springback for rail-type structures remains a significant challenge for automotive stamping. A major characteristic of forming such parts is that metals go through drawbeads and die-entry radii and often end up within part geometry (i.e., inside trimline). The bending-unbending stresses generated by drawbeads contribute significantly to the eventual springback. In production springback simulation, line drawbead models are generally used to represent the restraining forces provided by the real drawbeads for computational efficiency. While such models can be well correlated to match overall deformation of the part, the bending stresses could not be accurately captured. In the present study, a model of an aluminum U-channel is used to evaluate springback predictability of line drawbead model, which is then compared against simulations that employ detailed drawbead geometry. The results show that the line drawbead model largely under-predicts the springback and the bending moment. The accuracy of the prediction cannot be improved through different binder simulation strategies such as displacement-control or force-control. The study suggests that either real drawbeads be modeled, or the bending stress be incorporated into the line model to improve springback prediction.Copyright © 2006 by ASME