Prediction of Pollutant Emissions from Industrial Furnaces Using Large Eddy Simulation

摘要

Accurate prediction of pollutant emissions from turbulent combustion around complex geometries is of great practical interest. Here, an industrial furnace with rich-burn/quick-quench/lean-bum combustion for NO_x reduction is simulated. Large eddy simulation (LES) is employed as the simulation tool since it outperforms Reynolds-Averaged-Navier-Stokes (RANS) simulations in capturing large-scale unsteady motions and turbulent mixing. A structured LES code with variable staggering in space and time is used for accuracy and efficiency to predict NO_x and CO emissions from the furnace with a cross-flow-jet combustion system. The complex geometry of the furnace is handled by the Immersed boundary (IB) method. A Lagrangian dynamic subfilter model, which is designed for inhomogeneous flow in complex configurations, is employed. Laminar flamelet models are used for subfilter combustion closure. Here, three flamelet models, namely, the standard steady flamelet model, the Flamelet/Progress variable method, and the Lagrangian unsteady flamelet model, are employed. Experimental data obtained from the furnace with propane diffusion burners are used to compare the performance of these three models. Conclusions are drawn regarding the use of these models in combustion LES.