Numerical Analysis of Slag Transfer in the IronArc Process

The IronArc process is a novel approach to ironmaking which aims to reduce the associated $${\hbox {CO}}_{2}$$ CO 2 emissions. By superheating gas using electricity in a plasma generator (PG) the heat required for the process can be supplied without burning of coke. Reduction of hematite and magnetite ores is facilitated by additions of hydrocarbons from liquid natural gas (LNG). The melting and reduction of ore will produce a molten slag containing 90 pct wüstite, which will be corrosive to most refractory materials. A freeze-lining can prevent refractory wear by separating the molten slag from the refractory. This approach is evaluated in CFD simulations by studying the liquid flow and solidification of the slag using the enthalpy–porosity model in two different slag transfer designs. It was found that a fast moving slag causes a high near-wall turbulence, which prevents solidification in the affected areas. The RSM turbulence model was verified against published experimental research on solidification in flows. It was found to accurately predict the freeze-lining thickness when a steady state was reached, but with lacking accuracy for predicting the time required for formation of said freeze-lining. The results were similar when the $$k{-}\omega $$ k - ω SST model was used. A design with a slower flow causes more solidified material on the walls and can protect all areas of the refractory wall from the corrosive slag. A parameter study was done on the effect of viscosity, mushy zone parameter, heat conductivity and mass flow on the amount of solidified material, thickness of solidified material, heat flux, and wall shear stress. In the current geometry, freeze-linings completely protect the refractory for mass flow rates of up to 3 $${\text {kg}} \, {\text {s}}^{-1},$$ kg s - 1 , and are stable for the expected viscosity (0.05 to 0.3 Pa), heat conductivity (2 $${\text {W}}\, {\text {m}}^{-1}\,{\text {K}}^{-1}),$$ W m - 1 K - 1 ) , and used mushy zone parameter (10,000).