About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Phase Stability, Phase Transformations, and Reactive Phase Formation in Electronic Materials XXV
|
| Presentation Title |
Modeling Yield Strength of 7-series Aluminum Alloys: combining machine learning and physics-based methods |
| Author(s) |
Chien-Shuo Huang, Yu-Chen Liu |
| On-Site Speaker (Planned) |
Chien-Shuo Huang |
| Abstract Scope |
A machine learning model was developed to predict the yield strength of 7xxx series aluminum alloys by utilizing a comprehensive dataset consisting of alloy compositions, heat treatment processes, and corresponding mechanical properties. To ensure the model's physical relevance rather than being a mere statistical fit, we compared its predictions with those generated by the CALPHAD method. Furthermore, we translated the model outputs through precipitation-strengthening physical models across both under-aged and over-aged conditions. The results show that the model not only yields predictions that more closely match experimental data compared to conventional simulation tools, but also accurately reflects the microstructural evolution and underlying strengthening mechanisms. This demonstrates that the proposed data-driven approach possesses both predictive accuracy and meaningful interpretability, offering a promising route for accelerating alloy design while maintaining physical fidelity. |
| Proceedings Inclusion? |
Planned: |