| About this Abstract |
| Meeting |
2024 TMS Annual Meeting & Exhibition
|
| Symposium
|
Dynamic Behavior of Materials X
|
| Presentation Title |
Hybrid EAM-RANN Potential for Binary Ti-Al Alloy |
| Author(s) |
Mashroor S. Nitol, Saryu Jindal Fensin, Micah Nichols, Doyl Dickel, Christopher Barrett |
| On-Site Speaker (Planned) |
Micah Nichols |
| Abstract Scope |
Phase transformations in the Ti-Al binary system play a crucial role in influencing dislocation mo- bility and plasticity, making them a focal point of study. While interatomic potentials offer valuable insights into material behavior, existing models exhibit limitations in accuracy. Classical poten- tials like the embedded atom method (EAM) and modified embedded atom method (MEAM) yield satisfactory outcomes for solid solutions in the hexagonal close-packed (HCP) phase but fall short in approximating stacking fault energies in intermetallics, restricting their applicability in plas- ticity investigations. Current machine learning (ML) potentials primarily address intermetallics, overlooking critical transitions like α-Ti to β-Ti or α-Ti to D019. This study introduces a hybrid EAM-RANN potential, merging EAM with rapid artificial neural networks (RANN) to accurately reproduce the Ti-Al binary phase diagram up to 50% aluminum concentration. Leveraging Monte Carlo simulations, the EAM-RANN model adeptly anticipates phase transitions from α-Ti to β-Ti and α-Ti to D019. |
| Proceedings Inclusion? |
Planned: |