|About this Abstract
||2018 TMS Annual Meeting & Exhibition
||Frontiers in Solidification Science and Engineering
||Multiscale Modeling of Dendritic Grain Structures by Coupling DNN-CA-FE Methods
||Romain Fleurisson, Gildas Guillemot, Charles-Andre Gandin
|On-Site Speaker (Planned)
Industries strive to find accurate and precise quantitative multiscale models for solidification microstructures. In the present contribution, this is achieved using a combination of three methods: Finite Element (FE), Cellular Automaton (CA) and Dendritic Needle Network (DNN).
Macroscopic description of heat and mass transfer is provided by the FE method. Conservations equations for volume average quantities are solved at the scale of the simulation domain. The envelopes of the grains are tracked by the CA approach. The time and space dependent growth kinetics of the envelopes is inspired by the DNN method. The velocities stem from mass balance consideration in the vicinity of the dendrite tips. Solute flow is solved using a FE method with automatic adaptive meshing.
The combined method is validated against analytical results. Various quantitative simulations demonstrate the capabilities of the model for growth competition among columnar grains, equiaxed growth, and the columnar to equiaxed transition.
||Planned: Supplemental Proceedings volume