Extraction of the underlying material response of pseudoelastic NiTi and its application in numerical simulations



Journal Title

Journal ISSN

Volume Title



In certain temperature regimes NiTi exhibits pseudoelasticity, meaning that after being loaded to strains of 6-7% it can return to its original configuration. This behavior is produced by the reversible solid-state phase transformation between the austenitic (A) and martensitic (M) phases. During isothermal tensile testing the response produces a closed hysteresis that traces two stress plateaus corresponding to localization and propagation of transformation front(s). Hallai and Kyriakides (2013) extracted the underlying up-down-up material response during the A [rightwards arrow symbol] M transformation from an experiment on a laminate composed of an unstable NiTi core and hardening facestrips. In these experiments, the laminates were plastically deformed to a strain of about 6%. To obtain the underlying response during the reverse M [rightwards arrow symbol] A transformation, the laminate must be reverse loaded back to zero, resulting in compressive forces in the hardening facestrips which ultimately lead to the laminate buckling. This thesis presents a new experimental setup to prevent buckling by laterally supporting the laminate during reverse loading. From this test, the complete underlying NiTi response is extracted and exhibits the expected softening branches during both the A [rightwards arrow symbol] M and M [rightwards arrow symbol] A transformations, with each branch having a Maxwell stress similar to the corresponding experimental plateau stress level. The full response is used to calibrate a custom constitutive model that produces a fit based completely on a measured response for the first time. Simulations of the isothermal tensile tests using this fit capture the measured response and localized deformation pattern to the greatest extent thus far. The fit is also used to conduct a parametric study on the effect the hardening facestrip thickness has on the overall laminate response, and possible changes to aid future users of this method are identified. The new method presented can replace the previously empirical model calibration method and enable more confident modeling of the unstable behavior of SMA structures through the use of measured data.


LCSH Subject Headings