Combustion LES Software For Improved Emissions Predictions of High Performance Gas Turbine Combustors

The development of new military gas turbine combustors that produce low emissions of CO, NOx, and unburned hydrocarbons (UHC) is a difficult challenge for designers. Simulation tools that predict emissions are needed to reduce the cost of producingimproved, low emissions combustor designs. In this SBIR, we propose to continue the development of combustion Large Eddy Simulation (LES) methods that can more accurately predict emissions compared to current steady-state CFD codes. In Phase I of thisSBIR, the feasibility of using LES to predict CO and NOx emissions in a practical gas turbine combustor was demonstrated. Combustion LES predictions of a Rolls-Royce production combustor were compared with experimental combustor exit data. LES producedunsteady turbulent structures that enhanced mixing compared to steady-state CFD analysis. LES, at high power conditions, had good agreement with exit emissions data. However, at low power, LES predictions of CO emissions were poor, due to simplificationsin the fuel and CO oxidation mechanisms. In Phase II, the LES code will be improved by incorporating a multi-step assumed PDF method that accounts for more detailed kinetics with turbulent mixing interactions. Tabulation methods will be implemented andtested for improved computational efficiency. The final code will be validated against benchmark experimental data, and then applied to the Rolls-Royce and potentially Pratt & Whitney JSF combustors. The combustion LES software developed in this SBIR willbe useful in the cost-effective design and analysis of liquid-fueled, high performance combustors. The ability to reduce emissions, avoid combustion-driven instability, and to investigate high-payoff ideas will be possible. The software will be usefulfor gas turbine manufacturers, burner and boiler manufacturers, chemical processing industry, and the automotive industry. The software will be licensed and supported by CFDRC.