Thermomechanics of phase transformation induced localization in NiTi tubes. Part II constitutive modeling and simulations

Abstract

In an isobaric test, typically used for establishing the transformation temperatures of SMAs, a specimen is taken through a cool/heat cycle at a prescribed stress level. The companion paper, Part I, demonstrated that such tests on NiTi tubes result in rather complex interactions between the helical bands of localized deformation induced by phase transformations, the associated latent heat, and the thermal exchange with the environment. These interactions were quantified using accurately controlled testing conditions and full-field diagnostics, and provide a challenging platform for evaluating analyses. In Part II this challenge is met by first extending the thermomechanical constitutive framework developed by our group to include variations of transformation stress, strain, and latent heat with temperature. Unique features of the model include: modeling the reversible phase transformations of NiTi through a single surface in the deviatoric stress–temperature space with the transformation strain and entropy as the internal variables; and use of softening to model the inhomogeneous deformation exhibited in tension. The model is calibrated to the isothermal results of Part I. The constitutive model is then incorporated into a fully coupled static displacement-thermally transient finite element analysis that is used to simulate the isobaric experiments on NiTi tubes of Part I over a range of stresses. The clamped ends of the experiment are idealized as radial constraints. Heat exchange between the structure and the environment is strictly by convection. Isobaric testing is simulated by taking the model tube through the cool/heat cycle of the experiment at the prescribed stress level. A small thickness depression at one of the ends is used to initiate localized transformation. The temperature-strain response is reproduced with the two transformations initiating at essentially the same temperatures as in the experiments, producing the correct transformation strains for all stress levels. Transformation of M propagates via a helical band at similar speeds as in the experiments, while transformation of A is via multipronged fronts for all cases. Successful reproduction of the velocities of the banded transformations is governed by the value of the convection coefficient. Overall, the reproduction of the isobaric experiments over a range of stress levels, validates the constitutive and structural models. It also points to the limitations of modeling the heat exchange between the specimen and the environment only by convection.