Transformation Temperature Predictions Through Computational Intelligence for NiTi-Based Shape Memory Alloys

Abstract

Shape memory alloys (SMAs) are a class of metallic compounds that can return to their original forms, shapes, or sizes, when subjected to environmental stimuli, such as temperature and magnetic fields. Due to their unique shape memory effects and pseudoelasticity, SMAs, particularly NiTi-based ones, are of great interest in structures and composites, electronics, automobiles, biomedicine, and robotics. To tailor phase transformation temperature for practical applications, chemical substitutions have been extensively investigated and utilized. However, with multiple elements substituting for Ni, the correlation between the composition and transformation temperature is not elucidated but only the general trend is revealed with limited doping situations. In this study, we develop the Gaussian process regression model to find statistical correlations between NiTi-based SMAs’ transformation temperature ( $${T_{p}}$$ ) upon heating and nine pertinent physical parameters of alloying elements. More than 50 samples, with Ni partially substituted by one to three elements, are explored for this purpose. The modeling approach shows a high degree of stability and accuracy that contributes to low-cost $${T_{p}}$$ estimations.