CFD modelling of heat transfer in supersonic nozzles for magnesium production

An important step in the development of new processes is scaling up between laboratory operation, proof of concept pilot scale and final commercial scale plant. Heat transfer is one area where scaling is non-trivial and computational fluid dynamics provides one means of predicting scaling effects and adding understanding to the design process. CSIRO Minerals is developing a process for carbothermic reduction of magnesia and as part of that work CFD modeling of the process has been undertaken. This work outlines some of the CFD validation of heat transfer predictions between a supersonic gas stream and the nozzle wall and the impingement of the supersonic jet issuing from the nozzle on to a collector plate arranged normal to the flow. Heat transfer to the nozzle wall is well predicted by the k- turbulence model. Predictions of heat transfer for the impinging jet are less satisfactory with significant variation in heat flux predictions with changes in geometry, mesh resolution and turbulence model. For the later case the shear stress transport (SST) model gave the best predictions of heat flux and Nusselt number.