Abstract
An experimental campaign on some isotropic steels for pipeline applications has been put forth. It was based on tests with different stress states: tension on smooth and notched geometries, torsion, three point bending, plane strain, and combined tension-torsion. The aim was the characterization of the material elasto-plastic behavior up to large strain and the calibration of a ductile damage model for failure estimation. Results from tension and torsion were to be used for plasticity behavior description: from both of them, it is possible to retrieve the material true-stress true-strain curve until final failure. Unexpected differences were found. A critical interpretation led to the hypothesis that the ordinary isotropic J2-plasticity modelization could not predict all experimental evidences, starting from medium deformations, with increasing errors when plastic strain builds up. To approach this issue, an enhanced plasticity model has been developed and implemented into FEM code. It is a modification of a formulation widely used for geomaterials, which can take into account the influence of triaxiality and deviatoric effects on plastic accumulation. In addition, it allows a progressive transition from the Von Mises criterion, which demonstrated to be accurate for small strains, to a more complex formulation. A preliminary calibration of the model has been provided. This has been accomplished using a multiple-target inverse approach, exploiting experimental global data and FEA, by means of a dedicated optimization procedure. Appreciable improvements have been observed in terms of experimental-numerical match. Keywords Ductile materials • Mechanical testing • Large strains • J2-J3 plasticity model • Inverse calibration 35.1 Introduction The J2-plasticity is the most used theory for ductile isotropic materials; as its name suggests, it assumes that yielding and flow stress are governed by the second deviatoric stress invariant only. Limits of J2-plasticity are well-known since the works of Lode [1], Ros and Eichinger [2], Taylor and Quinney [3], who demonstrated that the model was not able to accurately reproduce all experimental evidences. Nevertheless, the correlation with experimental outcomes was found to be sufficiently good, in particular for the low-medium plastic range, so that the J2-plasticity, also thanks to its ease of calibration and implementation, has been widely employed for research purposes and industrial practice. The use of more comprehensive theories has been for a long time considered to add just more difficulties than benefits [4]. Nowadays, due to the improved ductility shown by materials, the correct identification of the plastic behavior up to very large strains has become a compelling instance, particularly for engineering applications where the modeling of stress and strain distribution at critical points, originated from complex loading conditions, is needed. Moreover, also ductile damage accumulation strongly depends on the state of stress, so that plasticity issues cannot be neglected for a good assessment of ultimate resistance of materials. It is then no coincidence that, in recent years, attempts to overcome J2-plasticity drawbacks have come from the ductile damage community.