| About this Abstract |
| Meeting |
TMS Specialty Congress 2025
|
| Symposium
|
3rd World Congress on Artificial Intelligence in Materials & Manufacturing (AIM 2025)
|
| Presentation Title |
Advancing Data-Driven Uncertainties and Predictions of the Stress Strain Response of Polycrystalline Alloys |
| Author(s) |
Jing Luo, Yejun Gu, Jaafar El-Awady |
| On-Site Speaker (Planned) |
Jing Luo |
| Abstract Scope |
Autonomous materials design requires reliable constitutive models to predict mechanical behavior efficiently and accurately. However, predicting the stress-strain response of polycrystalline metals remains challenging due to the complex interplay of microstructural features at multiple length-scales. Here, we present a physics-informed machine learning framework for predicting stress-strain curves of polycrystals by leveraging extensive experimental datasets from literature. Our approach is distinguished by incorporating the evolution of dislocation density as a key state variable governing plastic deformation, enabling physically meaningful predictions. Through Bayesian inference and uncertainty propagation methods, we develop a rigorous framework for quantifying uncertainties in the predicted mechanical response of polycrystals. Our models demonstrate robust extrapolation capabilities when applied to multi-principal element alloys through the integration of physics-based solution enhancement. Additionally, we analyze the fundamental differences in deformation mechanisms between face-centered cubic and body-centered cubic polycrystals, highlighting how crystal structure, deformation modes, and experimental data quality influence the model predictions. |
| Proceedings Inclusion? |
Undecided |