Advancing Nitinol Implant Design and Simulation through Data-driven Methodologies

A review and a perspective

Abstract

Recent advances in the Data Science methods for acquiring and analyzing large amounts of materials deformation data have the potential to tremendously benefit Nitinol (Nickel-Titanium shape memory alloy) implant design and simulation. We review some of these data-driven methodologies and provide a perspective on adapting these techniques to Nitinol design and simulation through a three-tiered approach. The methods in the first tier relate to data acquisition. We review methods for acquiring full-field deformation data from implants and methods for quantifying uncertainty in such data. The second tier methods relate to combining data from multiple sources to gain a holistic understanding of complex deformation phenomena such as fatigue. Methods in the third tier relate to making data-driven simulation of the deformation response of Nitinol. A wide adaption of these methods by the cardiovascular implant community may be facilitated by building consensus on best practices and open exchange of computational tools.