Prediction and reduction of springback in 3D hat shape forming of AHSS

Abstract

The overall objective of the current project was to develop a FE (Finite Element) model which can reliably predict springback in a 3-dimensional hat shape forming process. In this process beside wall angle change and sidewall curl, torsion or twist is also observed. The material used in this study is 1 mm DP980 and the material properties are obtained from IFU- Stuttgart. Yoshida model, as one of the advanced material models for accurate springback prediction, is used. A new approach called “inverse analysis” is also involved to evaluate the accuracy of springback prediction. Results show that the Yoshida model improves the springback prediction in terms of sidewall curl and wall angle change when 300 kN blank holder force is used. However, when applying higher blank holder forces (1500 kN in this case), simulation results with Hill48 and isotropic hardening are closer to experimental results. Inverse analysis can improve the simulation results in general. However, using the E-modulus which can provide accurate wall angle change does not improve the prediction of sidewall curl. Thus, in case of 3-dimensional U stretch-bending process, a constant average E-modulus is not appropriate to predict all three springback phenomena, i.e. wall angle change, sidewall curl, and twist. In agreement with literature, it is shown that additional stretching on the part can reduce springback. Also, simulations showed that using variable blank holder force, in the range considered in this study (from 300 to 1500kN), did not significantly reduce sidewall curl and wall angle change but it significantly reduced twist.