Evaluation of Subgrid Combustion Models at High Convective Mach Numbers

Abstract

The performance of LES subgrid scale turbulent combustion models were compared for a high speed reacting shear flow. This study considered a shear layer configuration with a convective Mach number of 1.3 and an adiabatic flame temperature rise of 985 K. Results for the assumed probability density function (pdf) and eddy dissipation concept (EDC) models were compared with results when the subgrid scalar fluctuations were neglected. The EDC model was found to produce unrealistic results as an apparent consequence of the limiting of product formation based on the subgrid turbulent kinetic energy. The pdf model resulted in a slight reduction of the mean product formation and heat release. lication of Reynolds Average Navier-Stokes (RANS) methodology to the simulation of high speed reacting ws suffers from a lack of validation data regarding turbulent transport properties (e.g., turbulent Prandtl and t numbers) and scalar fluctuations. In this flow regime, data of sufficient detail for model validation is difficult to obtain experimentally due to the challenging flow environment. One alternative avenue to obtain the required data to support RANS model development is through the use of large-eddy simulation (LES). LES could be used to generate the necessary data concerning fundamental flows of interest in a flight regime where experimental investigations are difficult or expensive to carry out. Also, LES data from a series of simulations spanning a wide range of flow conditions would support the development of physical understanding that will enable more universally applicable RANS models to be developed. This is important from the perspective that high speed aero-propulsive flows typically include mixed high and low speed regions, with the low speed regions potentially providing long enough residence times to influence the ignition and flame stabilization characteristics of a combustion device. A Following this philosophy of using LES to generate RANS validation data for the high speed flow regime, previous LES studies have been carried out for both nonreacting 1 and reacting 2 high speed shear layers. These studies compared high and low speed shear layers to determine the effect of compressibility on scalar fluctuations and turbulent transport properties. These studies lead to a number of interesting conclusions. First, compressibility effects were found to dramatically reduce the magnitude of scalar fluctuations compared to low speed flows. Second, compressibility effects were also found to modify the distribution of turbulent Prandtl and Schmidt numbers from a nonuniform profile at low speeds to an approximately uniform profile at high speeds. Third, heat release effects were found to decrease the asymptotic layer growth rate at low speeds but increase the growth rate at high speeds. For high speeds, the increased growth rate was likely due to the excitement of outer mode instabilities 2 . Fifth, shear layer mixing efficiency was found to mildly increase with increasing compressibility, suggesting that fluid unmixedness effects are less important for high speed flows than for low speed flows.