How modeling choices can affect the results of Ni-Ti stent fatigue assessment through surrogate specimens

Ni-Ti stents are the standard for the treatment of peripheral diseases; however, the reported failure in clinics highlights the issue of their fatigue characterization. One of the most common approaches consists of the experimental testing of surrogate specimens. On one hand, the use of “tension” specimens allows a straightforward strain evaluation, however, being quite different from the final strain distribution in the device. On the other hand, “bending” specimens replicate the strain distributions of the final device but in simplified and more easily tested geometries. This last strategy requires computational models to determine the local values and interpret the experimental results to identify the failure strain limit. These results are affected by the mesh refinement; indeed, different meshes result in different couples of mean and amplitude strains, also affecting the prediction of the location of the most critical region. Similarly, another overlooked critical choice is the element formulation: many studies testify to the absence of a consensus, even if in a bending-dominated problem it affects the distribution and maximum value of the deformation. The use of linear reduced elements enriched by a layer of membrane elements can increase the accuracy of the results. Given the literature gap on this topic, this work aims to investigate the role of different choices in the model preparation on the output of the computational fatigue analysis, performed through both tension and bending specimens. The same specimens used in the computational analysis are manufactured and experimental activity is performed to validate the findings.