Ratcheting in cyclic plasticity of metals: Experiments and modeling

The experimental study involved series of uniaxial and biaxial ratcheting experiments. Two types of biaxial experiments performed using thin-walled tube specimens. The first type involved axial, symmetric strain-controlled cycles carried out at constant internal pressure. This results in ratcheting in the circumferential direction. The second type involved axial stress-controlled cycles combined with constant internal pressure. Here, both the circumferential and axial strains ratchet. For each of the three loading types mentioned, a systematic set of experiments were performed from which the dependence of the rate of ratcheting on the experimental parameters was established.; The study was performed in two parts. In the first part, an effort was made to uncouple ratcheting from cyclic hardening or softening, in order to understand the ratcheting mechanism independently. To achieve this, the test specimens were cyclically stabilized prior to the ratcheting experiments. In the second part of the study, the complex interaction of ratcheting and cyclic hardening or softening was investigated.; Three rate independent cyclic plasticity models were critically reviewed with respect to the simulation of these ratcheting experiments. Their strengths, weaknesses and plausible alternatives are critically presented. The models considered cannot simulate the uniaxial ratcheting consistently for the experiments conducted. Using observations from the experimental results two schemes for alleviating this deficiency in the case of the Dafalias-Popov model are proposed in which the linear bounds are allowed to "shift" or "relax" with ratcheting. With these modifications, the model is shown to successfully reproduce all the uniaxial ratcheting experiments with good accuracy.; Prediction of the correct rate of ratcheting in the first set of biaxial experiments was found to be very sensitive to the yield surface kinematic hardening rule incorporated in the models. A hardening rule proposed by Armstrong-Frederick was found to provide reasonably good predictions. Prediction of the second set of experiments was found to require modeling of the changes induced to the shape of the hysteresis loops by this biaxial cyclic history. In the case of cyclically stabilized metals, the predictions are in the main of good quality. The corresponding predictions were found to be rather poor in case of cyclically hardening and softening materials. (Abstract shortened by UMI.)...