About this Abstract |
Meeting |
2024 TMS Annual Meeting & Exhibition
|
Symposium
|
Algorithm Development in Materials Science and Engineering
|
Presentation Title |
A Critical and Quantitative Comparison of Models for Grain Structure Prediction in Solidification Processes |
Author(s) |
Damien Tourret, Angela Montanero Lancharro, Rouhollah Tavakoli, Seyed Mohammad Elahi, Ignacio Romero |
On-Site Speaker (Planned) |
Damien Tourret |
Abstract Scope |
Prediction of solidification microstructure typically relies on either (1) accurate but costly models based on fundamental thermodynamics and kinetics, e.g. phase-field (PF), or (2) approximate but efficient models, e.g. based on Cellular Automata (CA). The discretization required for quantitative PF predictions severely limits their scale of application. In contrast, CA permits simulating entire 3D microstructures at process-relevant scales, but relies on strong assumptions, and how to choose a CA grid size remains unclear. Here, we quantitatively compare microstructures predicted by PF and CA at full melt pool scale. We show that, while CA cannot capture fine morphological details of grain boundaries, a reasonable agreement may be reached on the statistical distribution of grain orientations. We compare different CA algorithms to highlight quantitative discrepancies when high temperature gradients are present, and also discuss perspectives on Bayesian calibration of critical (yet poorly known) model parameters related to nucleation. |
Proceedings Inclusion? |
Planned: |
Keywords |
Modeling and Simulation, Solidification, Computational Materials Science & Engineering |