Simulation von Materialversagen an Gußbauteilen mit Hilfe eines duktilen Schädigungsmodells

The crash simulation of vehicles must incorporate the failure mechanism of materials, if the ductility of the material is so small that cracks can occur. In this paper we use the GURSON-model, a micromechanical damage formulation. The GURSON-model describes the damage mechanism as nucleation, growth and coalescence of microvoids. With the porosity, an additional internal variable, we have a measure for the crack initiaton. The model is used in the modified version of TVERGAARD AND NEEDLEMAN. An important aspect is the strong mesh dependence of the finite-element results. Reason is the strain-softening behavior of the GURSON-material. It is known that nonlocal rgularization methods must be used in such cases, when the governing equations change the type and lose ellipticity. But nonlocal formulations have the disadvantage that very fine meshes have to be used which is not convenient for explicit crash simulations where the elements should be big enough to ensure a reason able computing time. Therefore, the idea was here to introduce a functional dependence of the damage parameters from the absolute element size. The functional dependence has to be determined by the comparison of experimental data with different mesh size simulations. The modified GURSON-model was then tested by the comparison of the simulation with experimental Alumminum B-pillar tests where some cracks occur if the B-pillar is deformed extremly. The results show that the quality of the model is very good. It was possible to predict the crack locations and the crack sizes.