Search

Computational Fluid Dynamics has become a very powerful tool for developing engineering combustion devices, such as burners and furnaces. However, there are a wide variety of turbulence models, and some of them have proven to be more effective for some turbulent flow configurations than others. A reacting turbulent jet is a common flow configuration found in combustion engineering devices like burners. The present work assesses Reynolds-Average Navier-Stokes turbulence models, being tested on a CH4/H2/N2-Air reacting jet. Eight two-equation eddy-viscosity and three five-equation turbulence models were tested in the studied turbulent flow. Computational results were compared against experimental measurements in terms of flow field variables, mean mixture fraction, temperature, and species mass fraction. The findings suggest a strong influence of the turbulence model perforce on the mean mixture fraction as well as on the turbulence-chemistry interaction model. The modified version of the standard k-ε model proves to be the more reliable choice for this reactive flow configurations. Specially, where the flow patterns of the jet dictate the general flow physics. Near the fuel nozzle, both the Reynolds stress model with stress baseline k-ω (RSM-SBSL) and the standard k-ω model exhibit better agreement with experimental data than the conventional modified k-ε model. Moreover, findings from the standard modified k-ε model indicate a significant underestimation of spreading rates for radial samples in regions where jet spreading intensifies.

Large Eddy Simulation (LES) has become a powerful tool for dealing with turbulence. Nevertheless, mesh resolution of the simulated domain under the LES approach was shown to be a key matter. Critical regions should maintain an adequate mesh resolution and also the highest possible quality. LES with the Dynamic Smagorisky-Lilly sub-grid model was used for the simulation of SMA1-flame, to assess the influence of the mesh resolution on the instantaneous velocity fields, species profiles, and temperatures profiles. The results of the simulation suggest a comprehensible agreement with experimental data. Nevertheless, some areas in the highly rotational velocity field are not properly solved, mainly due to a poor mesh resolution of such areas. As a result, the central jet’s decay rate was not accurately predicted. On the other hand, the temperature and species profiles were reasonably computed, considering the simple chemistry provided by the Eddy Dissipation Model (EDM).