Comparing two selection laws of active slip systems in finite element polycrystalline model for numerical material testing

To meet the demand for high accuracy in metal forming simulation including difficult problems such as anisotropy, many material models have been developed. Since the recent material models usually possess many parameters and require cumbersome experiments, a reliable numerical material testing would be helpful to reduce the number of experiments. Therefore, we have engaged in development of a numerical material testing based on the finite element polycrystalline model in which the successive integration method is used for modeling slip systems. However, implementation based on the strain-rate dependent model, which is considered as the mainstream of such model, has not been rigorously considered in our research. In this study, two polycrystalline models were compared to establish better microstructural modeling for constructing a scheme of numerical material testing to predict material behavior that is not obtained by experiments. Numerical rolling, uniaxial tensile tests were conducted on aluminum alloy sheet with the strain-rate dependent model and the successive integration method. The crystal orientation calculated by the successive integration method exhibited close agreement with the experimental value of the rolled aluminum alloy sheet. On the other hand, the calculated crystal orientation by the strain-rate dependent model exhibited less close agreement with the experimental value of the same material than the successive integration method. To ascertain the characteristics of each model in terms of slip deformation quantitatively, the other tensile tests were conducted to calculate Lankford values caused by crystal orientation. Lankford values, calculated by the successive integration method, exhibited better agreement with experimental values than the strain-rate dependent model. These comparisons indicate that the successive integration method represented slip deformation more physically valid than the strain-rate dependent model and resulted in better calculation.