In earlier contributions, we discussed continuous-bending-under-tension (CBT) experiments on AA6022-T4. We found that CBT significantly enhanced the elongation-to-fracture and strength, over uniaxial tension. In the present paper, our understanding of CBT is expanded beyond these experimental observations, with the aid of material modeling and numerical simulations of the process. Cyclic tension-compression experiments were performed on this material, using strain histories that are expected to replicate the loading during CBT, i.e., different combinations of constant strain amplitude and linearly increasing mean value, to failure. During these experiments, a limited but not negligible amount of kinematic hardening was discovered. Some of these experiments are used for calibration of a combined isotropic-kinematic hardening model, while the rest are used for experimental validation of the model. The modeling framework is based on a rate-independent, associated flow rule with the von Mises yield criterion as the plastic potential. Isotropic hardening is introduced by a simple, exponential-decay model of the growth of the yield surface with plastic deformation. Non-linear kinematic hardening is introduced by a 4-term, Chaboche-type model. The large strain hardening curve is identified by extrapolation, an approach that is validated later in the work and contrasted with alternative options. This material modeling framework is introduced in finite element models of the CBT process. The model is meshed with linear, reduced-integration elements, with 7 elements through the thickness. It is found that the numerical model reproduces the experimental force-displacement curve, including the succession of spikes and plateaus typical of CBT, very closely. The model also replicates the development of strain on the surface during CBT, and compares well with post-test strain measurements. After these validations, the model is used to probe the mechanics of the CBT process, e.g., the development of stress and strain through the thickness and per cycle, the location and onset of failure, as well as the failure angle, which in CBT differs from the localized neck angle found in a typical uniaxial tension experiment.
Read full abstract