LES Prediction of Combustor Emissions From a Practical Industrial Fuel Injector

Abstract

An axial-swirl, lean premixed fuel injector typical of stationary gas turbine combustors has been analyzed using Large Eddy Simulation (LES). The objective of the study was to evaluate the LES modeling approach for predicting emissions of CO and NOx at practical engine conditions (P = 13.6 atm, Tin = 734 K = 861°F) and over a range of natural gas-air equivalence ratios (0.42 to 0.58). Experimental data from a recent UTRC/DOE-NETL program was used to evaluate the model. The experimental tests found NOx emissions decreased significantly with a decrease in equivalence ratio while CO emissions decreased initially, but then increased at the leanest conditions. LES calculations were performed using a parallel (domain decomposition), pressure-based, unstructured-grid flow solver within the CFD-ACE+ commercial software. The LES software solves the general transport equations for mass, momentum, energy, and chemical species without assumption at the grid- and time-resolved scales of the flow, and models the turbulent mixing and chemistry below the locally resolved grid/time-scales. The Localized Dynamic subgrid Kinetic energy Model (LDKM) was used to model the unresolved turbulence and a 2-step assumed PDF method, with decoupled NOx, was used to model the unresolved turbulence-chemistry interactions. Parallel calculations on a cluster of 22 Linux-based PCs were carried out. It was shown that LES was able to accurately predict the CO and NOx at an equivalence ratio of 0.58, and at leaner equivalence ratios the model was able to give qualitative agreement with the measurements. Some inadequacies in the NOx chemistry at ultra lean conditions and the near-wall flow boundaries were observed.