Empirical Rule-Based Estimation of Transformation Temperatures in Multicomponent NiTi-Based Alloys

Empirical rule calculations, widely used for phase selection in high-entropy alloys (HEAs), offer valuable guidelines for designing multicomponent solid solutions. In this work, we demonstrate that these empirical approaches can be effectively extended to the design of multicomponent intermetallics undergoing phase transformations, playing a remarkable role in estimating transformation temperatures in complex alloy systems. To this end, we compiled a comprehensive dataset of hundreds of medium- and high-entropy NiTi-based alloys exhibiting B2–B19′ transformations to establish such guidelines, enabling targeted exploration of the vast compositional design space and accelerating the discovery of new shape memory alloys (SMAs). Moreover, as machine learning (ML) continues to advance alloy design, this study shows that integrating key empirical and compositional parameters can significantly enhance ML predictive accuracy, even when working with limited datasets.