Prediction of NOx Emissions in an Industrial Gas Turbine Combustor Using Large Eddy Simulation and Reactor Network Modeling

The present investigation concerns prediction of NOx emissions in a stationary gas turbine combustor using reactor network modeling approach. Here, the reactor network is constructed based on spatiotemporal distribution of mixture fraction upstream of the flame front and flow residence time in the flame volume. To such aim, large eddy simulation is carried out to evaluate mixture fraction distribution upstream of the flame front, while the residence time is calculated by using RANS simulations. Moreover, the flame front position and flame volume is discerned using both RANS simulations and ENERGICO software. Obtained results are validated against experimental data. Such validations show that present method can accurately predict NOx emissions in the dry low emissions combustor. In an attempt to enrich the present investigation, parametric studies are carried out to evaluate effects of fuel composition, chemical kinetics, flame location, and combustion regime on the NOx production paths. Obtained results reveal that flame position and fuel composition have the most considerable effects on the NOx emissions among the investigated parameters. Based on above investigations, the present combustor is modified to reduce the NOx emissions from 25 ppmvd to 15 ppmvd.