AN IMPROVED ISOTROPIC KINEMATIC HARDENING MODEL FOR MODERATE DEFORMATION METAL PLASTICITY

An experimental program has been conducted to produce a number of different types of tests on several materials. The types of tests include: tension tests, large strain compression tests, reverse loading tests, symmetric strain cyclic tests, unsymmetric strain cyclic tests (mean stress relaxation), unsymmetric stress cyclic tests (cyclic ratchetting), and large strain torsion tests. Materials which have been tested are single phase 1100-O aluminum and 316 stainless steel, and two-phase spheroidized 1020, 1045 and 1095 plain carbon steels. These experiments represent the first collection of such a complete data base for any material. Within the framework of a phenomenological rate-independent plasticity theory with combined isotropic and kinematic hardening, specific constitutive equations have been formulated to represent our experiments and other results in the literature. An important new feature of the constitutive model is that during a reversing event the isotropic component of deformation resistance can initially soften. This softening is intended to represent an underlying dynamic recovery of the dislocation substructure. This dynamic recovery is dependent on the direction of straining, and is in addition to the usual dynamic recovery operative for monotonically increasing strains. The simple model captures key features of small strain uniaxial cyclic behavior reasonably well. Symmetric strain cycling, unsymmetric strain cycling, unsymmetric stress cycling and the additional hardening due to non-proportional strain cycling in tension-torsion are all captured by the model. Although the model predicts a monotonic increase of both the shear stress and the normal compressive stress toward saturation values in monotonic large strain torsion tests, the predicted stress levels are not in very good agreement with experiment. A modification of the constitutive equations considering a simple model for the plastic spin does not significantly affect the predictions of the model for simple torsion.