Suitability of yield functions for the approximation of subsequent yield surfaces (Q1369812)

Many constitutive models in the plasticity of metals are based on the existence of a yield function, which is not only used to mark the elastic limit but also as a potential function for the plastic strain rate. Measurements of elastic limits show how the geometry of the yield function contour lines should change with further straining. Existing proposals for invariant formulations of the yield function which take into account isotropic, kinematic and formative hardening are tested against typical geometric forms of the yield surfaces. Even if no evaluation equations are specified for constitutive variables contained in the yield functions, best approximations of measured yield surfaces can be computed by optimization of a quality function. Here the authors demonstrate that most of representations of the yield function are not able to describe experimental results qualitatively. The numerical results show that yield functions are essentially of grade three in the deviatoric stresses. Nevertheless, the evolution of internal variables can be deduced adequately from approximations based on the corresponding measurements.