|About this Abstract
||2022 TMS Annual Meeting & Exhibition
||Advances in Multi-Principal Elements Alloys X
||Predicting fundamental properties of refractory multicomponent alloys using electronic descriptors and statistical learning
||Yong-Jie Hu, Christopher Tandoc, Liang Qi, Peter Liaw
|On-Site Speaker (Planned)
Optimizing chemistries of bcc refractory multicomponent alloys to achieve a synergy of high strength and low-temperature ductility requires reliable predictions of the correlated alloy properties across a vast compositional space. In this work, first-principles calculations were employed to predict several strength/ductility-related fundamental alloy properties, including lattice parameters, lattice distortions, unstable stacking fault energies, and surface energies, for 106 individual bcc refractory multicomponent alloys. With the calculation results and descriptors based on electronic structures of interatomic bonding, statistical learning models were developed to efficiently predict these fundamental alloy properties for arbitrary alloy compositions without the need of any further first-principles calculation. The developed statistical models enabled rapid and systematic search of potential alloy candidates that are intrinsically ductile and with high yield strengths in high-order multicomponent systems.
||Computational Materials Science & Engineering, High-Entropy Alloys, Machine Learning