| About this Abstract |
| Meeting |
2025 TMS Annual Meeting & Exhibition
|
| Symposium
|
Bridging Scale Gaps in Multiscale Materials Modeling in the Age of Artificial Intelligence
|
| Presentation Title |
Developing On-Demand, Highly Efficient Digital Twins with DFT Accuracy for Iterative Alloy Discovery Frameworks |
| Author(s) |
Doguhan Sariturk, Guillermo Vazquez Tovar, Daniel Sauceda, Raymundo Arróyave |
| On-Site Speaker (Planned) |
Doguhan Sariturk |
| Abstract Scope |
Universal machine learning (ML) potentials are innovative computational models that leverage the power of ML algorithms to predict the properties of materials across a wide range of conditions. Once developed, these universal ML potentials can significantly reduce the computational effort required for material property predictions, facilitating efficient exploration and optimization of new materials. The present study outlines the implementation of a real-time and on-demand digital twin capability of universal machine learning potentials for accurate prediction of alloy properties at the DFT level of scale and accuracy. The proposed framework exhibits a high degree of computational efficiency, making it an ideal candidate for integration in iterative experimental materials discovery workflows. By utilizing this approach, the study aims to enable a predictive framework to facilitate the rapid development of advanced materials in a variety of property prediction studies, including mechanical, phonon, and thermoelectric properties. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Computational Materials Science & Engineering, Modeling and Simulation, Machine Learning |