Assessment of RANS-Based Turbulent Combustion Models for Prediction of Gas Turbine Emissions: Turbulence Model and Reaction Mechanism Effects

The goal of this study is to assess current, commonly applied turbulence and combustion models with respect to their performance in gas-turbine combustion (GTC). Reynolds Averaged Navier-Stokes (RANS)-based turbulence and chemistry models are two primary factors influencing the uncertainty in predicting turbulent combustion characteristics, especially for GTC. RANS-based methods are the design tools of choice in the gas turbine industry due to the high computational costs of LES (Large Eddy Simulation). In this study, lean premixed combustion of methane was simulated using two different reduced mechanisms (ARM9 and ARM19) along with the Eddy Dissipation Concept (EDC) turbulent chemistry interaction model to calculate the CO and NOx emissions. The effect of turbulence models was assessed by considering two different models. Both of the models tested performed well in the prediction of temperature and major species profiles. Predicted values of NO emission profiles showed an average difference of ±5 ppm compared to experimental values. Computed intermediate species profiles showed large qualitative and quantitative errors when compared with the experimental data. These discrepancies, especially the intermediate species hydrogen, indicate the challenges these reduced mechanisms and turbulence models can present when modeling pollutant emissions from gas turbine combustors. 1