Characterization and Validation of Fatigue Strains for Superelastic Nitinol Using Digital Image Correlation

Abstract

Fatigue is a major challenge encountered in cardiovascular implant design. While the properly heat-treated Nitinol can exhibit up to 6–7% recoverable strains allowing for minimally invasive transcatheter delivery of cardiovascular implants, the cyclic in vivo loading can cause premature fracture of the implant if the fatigue strain is too high. Strain-based criteria have been adopted for the development of Nitinol fatigue resistance. Lacking experimental tools to characterize the local material fatigue strain, fatigue testing of Nitinol specimens has largely relied on the finite element analysis to compute the cyclic strain amplitude and mean strain based on experimentally derived constitutive parameters using phenomenological strain energy theory. Without a consistent computational standard, previous works have resulted in controversy and inconsistency in the impact of mean strain on the fatigue resistance of Nitinol in terms of strain amplitude limit at high cycle fatigue regime. In this paper, digital image correlation (DIC) technique is used to experimentally determine local material strains of Nitinol fatigue specimens using monotonic and cyclic loading conditions. These local strains are compared with strains computed from finite element analysis. It was found that strains from DIC and FEA are comparable in the single-phase states (pure austenitic or martensitic), whereas the measured strains can show significant difference from simulation computed strain during the transformation stage where both austenite and martensite phase co-exist. These observations have significant implications to nitinol fatigue testing and implant reliability assessment.