Assessment of RANS-based CFD Approaches for Numerical Modeling of Bluff-body Diffusion Flame Burners

In the present work, a systematic numerical validation study is carried out to assess the capabilities and limitations of Reynolds-averaged Navier-Stokes (RANS)-based computational fluid dynamics (CFD) approaches for modeling of methane-hydrogen (CH4-H2) bluff-body diffusion flames. Both non-swirling and swirling diffusion flame configurations, corresponding to the National Energy Technology Laboratory (NETL) Enclosed Sydney Swirl Burner (ESSB), are considered for a 50%:50% (by volume) CH4-H2 fuel blend. Various turbulence models, combustion models, and chemical kinetic mechanisms are tested to assess their predictive accuracy. For the non-swirling flame, Reynolds stress model (RSM), in combination with the flamelet generated manifold (FGM) model and FFCM2 mechanism, is found to provide the best prediction for flame structure and velocity profiles, when compared with NETL-ESSB experimental data. For the swirling flame, however, the same CFD model underpredicts swirl number. The model sensitivities to geometric effects and boundary conditions are further investigated with the goal of mitigating this discrepancy. Results show that by scaling up the inlet tangential velocity by 20%, improved prediction of the flame structure and velocity fields can be achieved for the swirling flame. It is also observed that while both FGM and finite-rate chemistry (FRC) models predict similar flame structure for the swirling case, FRC shows much higher hydroxyl (OH) mass fraction in the flame compared to FGM model. Overall, the FGM model shows a better performance across both the non-swirling and swirling cases in the RANS framework.