| About this Abstract |
| Meeting |
2022 TMS Annual Meeting & Exhibition
|
| Symposium
|
Grain Boundaries and Interfaces: Metastability, Disorder, and Non-Equilibrium Behavior
|
| Presentation Title |
Machine Learning-assisted Prediction of Interfacial Segregation in a Refractory Multi-principal Element Alloy |
| Author(s) |
Doruk Aksoy, Timothy J. Rupert |
| On-Site Speaker (Planned) |
Doruk Aksoy |
| Abstract Scope |
The evolution of models explaining segregation behavior of solutes continues at a rapid pace. A complexity that requires elucidation is the co-segregation behavior in multi-principal element alloys (MPEAs). Co-segregation depends not only on the chemical ordering in the MPEA, but also on the interface structure and its evolution during the segregation process. In this work, we categorize solute-solute interactions using dimensionality reduction techniques for a NbMoTaW refractory MPEA polycrystal. We investigate both the dilute limit and chemically complex environments for the solutes. The evolution between these states is also studied via hybrid Monte Carlo-Molecular Dynamics simulations. Finally, we utilize all the segregation energies and local atomic environment vectors and feed them into an artificial neural network algorithm to predict segregation behavior in the complex alloy. The diversity of structural and chemical motifs present in the analysis further validates the complexity of the problem and assist the interfacial design of MPEAs. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
High-Entropy Alloys, Modeling and Simulation, Machine Learning |