Evaluation of methods for determining the Yoshida-Uemori combined isotropic/kinematic hardening model parameters from tension-compression tests of advanced lightweighting materials

The performance of two methods, termed Manual and Optimization, for determining the values of the constitutive-model parameters for Yoshida-Uemori (Y-U) isotropic-kinematic hardening model was evaluated. The Y-U parameters were determined for two 6000-series aluminum alloys (AA6XXX-T4 and AA6XXX-T81) and two dual-phase steels (DP 980 and DP 1180) using stress-strain data from tension and tension-compression tests. In the Manual Method the parameters were calculated by sequentially fitting the forward and then the reverse segments of tension-compression stress-strain data. In the Optimization Method, the parameters were calculated by systematically reducing the difference between the data and stress-strain curves computed from a one-element finite element model coupled with the optimization module of the commercial finite-element software package. The performance of the methods was evaluated qualitatively by graphically comparing measured and computed stress-strain curves and quantitatively by evaluating the residual between the measured and computed curves. Although the quality of agreement for the Manual Method was generally good for the aluminum alloy tests, the agreement was inferior for the dual-phase steel tests, especially for the reverse deformation segment. The Optimization Method produced the best results as evaluated by the lowest residual between the measured and computed curves. Of several variants of the Optimization Method, simultaneous optimization against all three replicate curves produced Y-U parameters that best represented the measured stress-strain curves. Statistical analyses conducted in conjunction with optimization studies also established the relative influence for each of the Y-U parameters on the overall agreement between measured and computed curves. A broader implication is that using a finite element model, rather than the manual methods, should be used to determine the Y-U pa-rameters, since those parameters will be eventually used in a finite element model to simulate a forming process. This is especially true when determining a single set of parameters for repeat tests.