The development of numerical models for reacting flows has seen significant progress in recent years. However, due to the high complexity of the interaction between turbulence and chemistry in different combustion regimes (premixed, partially premixed and diffusion controlled), many scientific questions are still open. Even within a single regime, various turbulence-chemistry interaction mechanisms lead to different flame behavior. Therefore, it still make more sense to describe each particular regime by its own specific combustion model.

The aim of this work is the evaluation of the effect of subgrid scale modeling, validation and comparison of different combustion models in context of Large Eddy Simulation (LES) for premixed flames. The in-house code FASTEST is used for the numerical investigation of the premixed combustion in the turbulent flames with gas turbines characteristics, i.e., high turbulence intensity and high energy density. In the following the developments and results based on the current work are summarized:

The current work introduces an extension of an artificially thickened flame approach with a dynamic wrinkling model. The dynamic power-law wrinkling model is coupled with the Flamelet Generated Manifolds (FGM) tabulated chemistry. To analysis this model improvement, simulations are conducted for a lean premixed Bunsen type flame (Matrix Burner) with high turbulent intensity, using the ATF (Artificial thickened flame) combustion model on two grid levels. For validation of results, the same simulations are conducted with other combustion model, F-TACLES (Filtered tabulated chemistry), both with dynamic and non-dynamic formulation of power-law wrinkling model. The dynamic formulation of the F-TACLES approach has been implemented in the same code and evaluated for different configurations The numerical investigations using non-dynamic version of wrinkling models coupled with both combustion models on all two grid levels fail to predict the flow and temperature characteristics. The simulation using developed dynamic power-law wrinkling model shows excellent results comparing to the experiment independent of grid level used. Series of simulations are conducted in the frame work of "combustion noise'' research project in collaboration with three other research groups. The newly developed dynamic power-law wrinkling model coupled with artificially thickened flame model and FGM chemistry is also applied to two chosen test cases. The effect of the dynamic model is clear to observe but limited as the flames are positioned in the quasi-laminar regime. The conclusion is that: for the correct prediction of premixed flames with high turbulent intensity, elaboration of combustion models coupled with dynamic subgrid scale modeling is inevitable.