About this Abstract |
Meeting |
2023 TMS Annual Meeting & Exhibition
|
Symposium
|
Deformation-induced Microstructural Evolution during Solid Phase Processing: Experimental and Computational Studies
|
Presentation Title |
Three-dimensional Phase-field Simulation of Static Recrystallization in Aluminum Alloy Combined with Bayesian Data Assimilation |
Author(s) |
Kota Matsumoto, Eisuke Miyoshi, Yoshiki Mori, Kishu Akiba, Masato Ito, Nobuhiro Kitahara, Kenichi Yaguchi, Akinori Yamanaka |
On-Site Speaker (Planned) |
Kota Matsumoto |
Abstract Scope |
Multi-phase-field (MPF) method is widely used for simulating microstructure evolution in static recrystallization of alloys. However, it is difficult to measure distribution of stored energy in deformed grains and calibrate model parameters (e.g., grain boundary energy) used for the MPF simulation, which determine the predictive accuracy of the simulation, only from experiments. In this study, we propose to employ Bayesian data assimilation based on the ensemble Kalman filter (EnKF) to estimate the model parameters and improving the simulation accuracy by integrating in-situ EBSD observation data into the simulation resutls. To validate the proposed methodoloy, we estimate the stored energy distribution in the deformed grains, the grain boundary energy and mobilitity by performing three-dimentional MPF simulation of static recrystallization occurred in cold-rolled industrial pure aluminum using the EnKF-based data assimilation method. |
Proceedings Inclusion? |
Planned: |
Keywords |
Aluminum, Machine Learning, Process Technology |