Numerical study of effect of surface roughness on superelastic behaviour of shape memory alloy

In present paper, a 3-D thermo-mechanical constitutive model for the polycrystalline shape memory alloy (SMA) has been used to predict the effect of surface roughness on superelastic behaviors. Periodic surface notches with four different notch depths as surface roughness were introduced into the models. For comparison, simulations were also performed on smooth models and models with a single notch. From the results, it is found that for singly notched models with small notch depths, the Mises equivalent stress distribution is similar to that of smooth model, with the increasing of notch depth, the maximum value of Mises stress level increases, and the maximum value of martensite average volume fraction becomes big. With the increasing of surface roughness, the maximum Mises stress level increases, and the stresses distribution becomes non-uniform. From the comparison between the singly notched models and rough surface models, it is clear that with the increasing of notch depth, the level of maximum Mises stress-axial strain curves lift, and the maximum value of averaged martensite volume fractions become big for singly notched models, but for rough models, it is almost similar to that of smooth model. Because of the interference between notches, when the depth of notches becomes big, the maximum value of Mises equivalent stresses for rough models is higher than that for the singly notched models even with same order of the notch depth.