Inverse method for identification of initial yield locus of sheet metals utilizing inhomogeneous deformation fields

Accurate finite element simulation of sheet metal forming processes requires among others accurate description of plastic behaviour of materials. This is achieved by utilization of sophisticated yield criteria having several material parameters. This work proposes a procedure which makes use of the distribution of strains to identify the initial yield locus of sheet metals by the help of inverse analysis. For this purpose a flat specimen having a varying cross-section is introduced, which is capable of revealing different deformation states in one test. Numerical simulations are performed with 2 representative materials for steel and aluminium, using the material model Yld2000-2d. The results of these simulations are treated as experimentally obtained results and with the inverse methods it is tried to obtain the given yield locus. The relation between the supplied input and the outcome of the inverse algorithm is studied by examining different objective function definitions. The numerical studies show that inclusion of the strain distribution in the definition of objective function is a key issue in identification of the yield locus. The orientation of the specimen with respect to the rolling direction also determines the amount and quality of the information used for parameter identification. Consequently the circumstances, under which the inverse method can predict the initial yield locus, are defined.