About this Abstract |
Meeting |
2023 TMS Annual Meeting & Exhibition
|
Symposium
|
AI/Data Informatics: Computational Model Development, Validation, and Uncertainty Quantification
|
Presentation Title |
Machine Learning Assisted Candidate Search For Niobium Alloy |
Author(s) |
Trupti Mohanty, K. S. Ravi Chandran, Taylor D. Sparks |
On-Site Speaker (Planned) |
Trupti Mohanty |
Abstract Scope |
Niobium-based alloys can be considered promising high-temperature turbine materials because of their exceptional properties at elevated temperatures. However, the conventional way of designing an optimal alloy composition using limited available experimental data and from a vast unexplored compositional space is a challenging task. In this study, we provided a machine learning-based design strategy that has been applied to a huge compositional space for finding suitable Nb-based alloy composition for the optimal value of yield strength (YS) and ultimate tensile strength (UTS). We have used the Bayesian optimization algorithm combined with domain knowledge-based material descriptors to propose an optimal Nb-based quaternary and quinary alloy composition. This design strategy has also been extended to simultaneous optimization of multiple targeted properties. In addition to optimal alloy composition selection, each alloying element's criticality, availability, and sustainability have also been assessed by analyzing the economic factor Herfindahl–Hirschman Index (HHI) based on its reserve and production. |
Proceedings Inclusion? |
Planned: |
Keywords |
Machine Learning, |