ASSESSMENT OF AN EQUIVALENT REACTION NETWORKS APPROACH FOR PREMIXED COMBUSTION

The pollutants produced by the burning of fossil fuels have a severe impact on the environment and on mankind. Computational fluid dynamics (CFD) is a powerful tool, which is widely used to predict the emission of these pollutants from industrial combustion systems. Nevertheless, to predict these emissions the chemical reaction must be represented by a detailed mechanism, which includes pollutant formation pathways. Thus, using a complex mechanism, especially in 3D simulations with a realistic geometry is prohibitively expensive computationally. In this article, the equivalent reaction networks (ERN) method is used in conjunction with a Reynolds-averaged Navier Stokes (RANS) approach to reduce the cost of these computations. For this purpose, a pilot stabilized stoichiometric methane-air flame is chosen with a specific interest in species with slow time scales, such as CO and NOx. The Favre averaged CFD results are then compared to previously-reported experimental measurements and earlier computations using conditional moment closure (CMC) at five axial locations within the flame. Despite the simplicity of the ERN method in contrast with other more complex combustion models, the comparison of the CFD results with the experimental measurements for the prediction of CO are extremely encouraging.