| About this Abstract |
| Meeting |
TMS Specialty Congress 2025
|
| Symposium
|
The 7th International Congress on 3D Materials Science (3DMS 2025)
|
| Presentation Title |
A Novel Method for Microstructure Prediction of Austenitic Stainless Steel Based on Machine Learning |
| Author(s) |
Yuqing Du, Jun Sun, Håkon W. Ånes, Fei Chen |
| On-Site Speaker (Planned) |
Yuqing Du |
| Abstract Scope |
The prediction of mixed grains is crucial for determining whether austenitic stainless steel achieves excellent mechanical properties after solution treatment. In this presentation, machine learning approach is adopted to predict the microstructure evolution of austenitic stainless steel during hot deformation and solution treatment. Five different machine learning models are compared, using experimental data from large-area EBSD measurements covering mixed grain areas. Shapley Additive exPlains (SHAP) method was found to be most accurate interpreting the model to find key factors affecting the formation of mixed grains in hot deformation and solid solution treatment. Moreover, a time series 3D grain data about these key factors was researched from lab-based diffraction contrast tomography (Lab-based DCT) characterizing the grain coarsening processing during solution treatment. Various aspects coupling 2D and 3D experimental data to machine learning will also be discussed. |
| Proceedings Inclusion? |
Undecided |