Towards Efficient Modelling of Slag Entrainment during Metallurgical Processes

This paper focuses on the numerical modelling of slag entrainment in liquid metal. Due to the complexity of the multi-phase flow situation and the inherently unsteady nature of the process, feasible simulations are restricted to rather coarse grids. Nevertheless, important flow structures like secondary suction vortices cannot be captured by an insufficiently resolved grid. As a consequence three numerical approaches are proposed in order to focus on local slag entrainment events within a global flow simulation. Firstly, in the course of Volume of Fluid (VOF) simulations the evolving metal-slag interface is sharpened by a solution dependent adaptive grid refinement. Secondly, possible suction vortices are tested by super-imposed Chimera grids. In this approach a finely resolved O-shaped grid is placed around the trajectory of a representative Lagrangian fluid particle that is started from a local depression at the metal-slag interface. Thirdly, a simplified concept of Lagrangian slag droplets is utilized in order to detect irregularly occurring flow situations that are prone to slag entrainment. These modelling approaches are applied to metallurgical processes like tundish pouring or continuous casting. With help of these slag entrainment modelling approaches a global metallurgical flow simulation can be augmented by the effect of local entrainment events.