A thorough understanding of multiphase flow phenomena is crucial for the optimization of metallurgical processes. This project focuses on modelling and numerical simulation of (i) slag entrainment, (ii) bubble induced recirculation in vacuum treatment plants and (iii) flotation of non-metallic inclusions.

The main objective of this project is to further develop numerical models for metallurgical flows enabling an in depth investigation and optimization of metallurgical processes such as (i) high quality continuous casting, (ii) steel refinement in RH-plants and (iii) particle separation in tundishes.

  • Numerical model development for industrial multiphase flows
  • CFD simulation of flotation, continuous casting, and RH-plants
  • Experimental validation of modelling methods and results


This model development project assembles two PhD works and additional postdoctoral work. In a first PhD work the phenomenon of slag entrainment at the steel-slag interface in continuous casting mould flow is considered. In setting up a simulation methodology this flow situation will be considered as a multi-scale challenge comprising a macroscopic flow pattern which might trigger microscopic flow events, i.e. slag entrainment, at the steel-slag interface. In the current research period we put forth a Lattice-Boltzmann (LB) based turbulence model for the detailed analysis of turbulence modulation in single phase flows. In the upcoming research period we will further develop this modelling concept for stratified fluids. While global mould flow pattern will be pictured by classical macroscopic flow simulations, small-scale interface instabilities will be resolved by an embedded LB based turbulence model. Numerical model developments within this PhD work will be validated by dedicated experiments available at JKU and RHI (see also Project 4.5).

In a second PhD work the recirculation flow inside a vacuum degassing plant, a so-called Ruhrstahl-Heraeus (RH) plant, is highlighted. From a fluid dynamical point of view the main challenge of this PhD work is predictive modelling of massive gas injection into a low pressure environment. Once the resulting steel recirculation flow is validated by surface velocity pattern obtained from on-plant video processing, additional heat transfer and chemical reaction phenomena are considered. In developing dedicated models for heterogeneous reactions the JKU (specialist in fluid dynamical modelling) will join forces with MUL (specialist in metallurgical reactions).

Finally, a third work package will focus on the attachment of non-metallic inclusions to uprising bubbles within a bubble swarm. This phenomenon will be studied both analytically and by means of comprehensive numerical simulations. To achieve this, an existing standard model for bubble plumes will be augmented by dedicated sub-models, which, in turn, are deduced from a combination of drift-flux theory for rising bubble plume and collision theory for the interaction between bubbles and particles. Finally, numerical results will be tested by plant data of argon curtains in steel tundishes.

Results and Application

The first PhD work in P4.4 Liquid Melt Models will provide an in depth understanding of slag entrainment event. Thereby, critical macroscopic flow pattern, being prone to slag entrainment, will be identified. These findings will directly feed model developments in project P4.5. Steel producer and manufacturer of continuous casters will get knowledge of critical flow pattern, which are prone to deteriorating the quality of the final product. Additionally, they will develop control methods, avoiding those flow situations in operating or designing continuous casters. The second PhD work on steel recirculation flows will provide a new tool for the optimization of operation of vacuum treatment plants. Steel producers will get a prediction tool for the overall heat loss and reaction turn-over of a melt during vacuum treatment in dependence on initial plant conditions and operating parameters. Manufacturers of vacuum treatment plants, in turn, will get a development tool for geometry optimization. Finally, the last work-package will deliver a prediction tool for the design and efficient operation of gas bubble curtains.