Failure model calibration of a DP1000 dual phase steel using solid and shell elements for crash simulation

In the trend of lightweight construction more and more advanced high strength steel sheets (AHSS) come to application in automotive structural components. Since the ductility of high strength steels is relatively low, damage behavior of these materials must be modeled. In automotive structures AHSS components are usually discretized with shell elements. With the increase of computation capacity attempts with solid elements are made to capture the loading state more accurately, especially after necking. For this reason, it is convenient to develop a method which enable a systematic model calibration from the 3D to the 2D loading situation. The failure strain of metallic materials depends on stress state. In the past years several studies have shown that the stress triaxiality is not sufficient to describe failure and empirical models were extended to consider the effect of Lode parameter. Recently it was also show that the amount of bending seems to influence failure, namely the failure strain increases with the amount of bending. In this work a DP1000 dual phase steel was investigated both experimentally and numerically. To study the dependence of damage behavior on loading state, smooth, notched and shear tensile tests and tension-bending, punch and Nakajima tests were performed. A general damage model based on a critical failure strain depending on triaxiality and Lode parameter was calibrated for solid elements. The failure stain curve for shell elements was derived from the failure surface for solid elements considering the relationship between triaxiality and Lode parameter under plane stress condition. Moreover, the influence of the bending factor was incorporated into the failure model for shell elements. FE simulations with LS-DYNA were performed to determine the local values of triaxiality, Lode parameter and bending ratio for each specimen type. The applied damage models were calibrated and verified by simulating all specimen tests. It is shown that the developed method well captures the failure behavior for solid and shell elements. The predictions were better using shell elements because of the possibility to account for the bending ratio influence, what is not possible using solid elements.