Inverse Identification Of Plastic Material Behavior Using Multi-Scale Virtual Experiments

Mixed numerical-experimental techniques used to identify plastic material properties of sheet metal are conventionally based on experimental data (e.g. full-field data) acquired during mechanical experiments. Although those techniques definitely enable to reduce the experimental effort for identifying plastic material properties, accurate identification of advanced phenomenological plasticity models still requires a significant amount of experimental effort. In this paper, we explore the opportunity to further reduce this experimental effort by replacing the mechanical experiments by virtual experiments using a physics-based multi-scale model. To this purpose, the Alamel polycrystal plasticity model, which solely requires the input of the initial crystallographic texture and a single tensile curve, is used to generate virtual plastic work contours in the first quadrant of stress space. The generated virtual experimental data is then used to inversely identify a phenomenological yield function. Finally, the predictive accuracy of the proposed method is investigated by using a finite element code to simulate the hydraulic bulge test.