論文ID: ISIJINT-2025-159
To quantitatively predict the microstructure using a cellular automaton (CA) model, problems such as the calculation accuracy of the interface curvature and occurrence of mass-balance errors must be overcome. In this study, we focus on mass-balance errors, develop a unique CA model that considers the solid–liquid interface movement associated with solidification and melting, and investigate the effects of mass-balance errors on microstructural evolution. Mass-balance errors occur in CA models based on a two-domain approach that separately defines the solute concentration distribution in solid and liquid phases. We identified the causes of mass-balance errors and quantitatively estimated them, and then conducted a simulation of the microstructural evolution of Al-4.5mass%Cu alloy during isothermal undercooling and continuous cooling with and without mass-balance error corrections. During isothermal undercooling, if the mass-balance error is not corrected, the solute concentration increases during the coarsening of the dendrite arms, and the increase is greater with increasing undercooling, which causes the solid–liquid interface to melt despite the isothermal process. During continuous cooling, if the mass-balance error is not corrected, the solute concentration decreases during arm coarsening in the middle of solidification. Under the simulation conditions, the absolute percentage error of the solute concentration relative to the initial composition exhibited maximum values of 1.31% and 14.9% during isothermal undercooling and continuous cooling, respectively. Under both conditions, the values are almost zero if the mass-balance error is corrected. Therefore, mass-balance errors can be appropriately corrected in the proposed CA model based on a two-domain approach.
