High Fidelity CFD Analysis of a Stationary Mixer

Abstract

In the process of cleaning ue gases by chemical reactions, such as with Selective Catalytic Reduction (SCR) cleaning, the performance of the separator device is among others dependent on the mixing of gases before the catalyst region. Hence, in order to develop new products or improve existing ones it is important to have efficient tools to predict the mixing. Previous work using CFD (Computational Fluid Dynamics) on mixing show that there is a need for a more accurate simulation tool than steady state RANS. This thesis presents a comparative study of the ability of di erent turbulence models to predict the molecular mixing within an inhomogeneous multi-species fluid mixture in the ow through a SCR duct test rig, for which data is available from laboratory testing. Simulations were carried out using ANSYS-CFX, comparing a steady state RANS model, an unsteady RANS model and a Scale-Adaptive Simulation model (SST-SAS). Three meshes of different densities were used. The models abilities to predict the pressure drop over mixers was also noted. Results showed that all models proved satisfactory in predicting the pressure drop over the mixers, with the highest accuracy when using SST-SAS on the finest mesh. It was evident that the RANS models underpredict the mixing whereas the SST-SAS provides results consistent with experimental data. Although the SST-SAS was more computationally demanding than the RANS models, it was concluded that it is more efficient to use turbulence models designed for resolving turbulent motion on a coarse mesh than using time averaging turbulence models on finer meshes.