MULTI-PHYSICS SIMULATION IN THE 21st CENTURY.

Email: info@physica.co.uk

[ ALUMINIUM REDUCTION CELL ]   [ CASTING & SOLIDIFICATION ]
[ ELECTRONICS APPLICATIONS ]   [ FORMING ]   [ GRANULAR FLOWS ]
[ STEEL PROCESSING ]   [ WELDING ]

Direct smelting

The direct smelting process involves the coal based direct smelting of iron ore to produce reduced iron units which may be immediately refined to steel. An example of this is the HIsmeltŪ Process which uses the Smelt Reduction Vessel (SRV) [ IMAGE ]. Within the SRV volatiles in the coal are cracked, carbon is dissolved in the metal, ore is smelted and molten slag formed. Since all the above processes are endothermic there is a requirement to provide a substantial amount of energy to maintain the bath temperature. The key to the direct smelting process is the transfer of energy, resulting from chemical reactions in the gaseous top space region, to the metal droplets and the bath.

Model

To facilitate the modelling the SRV is divided into three regions namely the Bath Zone, the Transition Zone and the Gas Zone [ IMAGE ]. The bath zone is the volume of the converter where the liquid and slag form a continuous phase, the gas zone is where the gas phase is continuous and the transition zone is where the gas phase is continuous but there exists a significant amount of liquid in the form of droplets.

The modelling performed using the research version of PHYSICA concerns the fluid flow and heat transfer process in the top space (The top space being the combined gas and transition zones). The major aspects of the model required to simulate the top space are:

• A turbulent reacting gas space model.
• Lagrangian particle tracking for the liquid droplet phase.
• Models for the transfer of heat and mass of all chemical species between the continuum and droplet phases.
• A radiation model to calculate radiant fluxes.
• A gas turbulence model for regions of high swirl.

Results

One of the topics of interest within the direct smelting process is the use of multiple swirling jets. The use of high swirl number jets results in reverse flow in the centre of the jet. This increases the surface area of the jet which promotes the reactions of the chemical species. As a consequence, there is a greater possibility of the energy produced by the reactions being transferred to the metal droplets and bath.

To enable the modelling of the twin-tuyere SRV algorithms had to added to PHYSICA to model the anisotropic nature of turbulence in regions of high swirl. The model employs the standard ke model for the calculation of the length scale to be used by the radial and axial components of velocity and a mixing length based model for the length scale in the tangential velocity direction. The difficulty with this model lies in the need to ascertain radial, axial, and tangential directions for the flow on an unstructured mesh. Because of the interaction of the jets it is not possible to assume that the flow axis is aligned with any mesh feature. Consequently, there is a need to calculate the position of the flow axes. Having evaluated these axes it is possible to calculate the relevant cylindrical directions for each element. An adjustment to the flow module was also made to allow correct interpolation of the velocity vector (In earlier simulations it had been shown that interpolation of the Cartesian components of velocity in swirling flow significantly over-predicted the decay of swirl).

The images below show the mesh and some of the results from the simulation carried out using PHYSICA:

Mesh [ IMAGE ], Axial velocity [ IMAGE ], Swirl velocity [ IMAGE ]

The mesh, generated in PATRAN, consist of 90,000 elements. The simulation employs the anisotropic turbulence model and a Simple Chemical Reaction Scheme (SCRS) to simulate the reaction of the species.

Axial velocity at varying heights up the SRV

Top [ IMAGE ], Middle [ IMAGE ], Bottom [ IMAGE ]

Chemical species

Oxidant [ IMAGE ], Fuel [ IMAGE ]