Modeling and uncertainty analysis of chemical reactor network model in lean premixed combustion chamber

The chemical reactor network (CRN) model based on three-dimensional numerical simulation results of CFD has great advantage in fast and accurate prediction of NOx emissions for combustor. In this work, the characteristic distributions of flow and temperature fields in a lean premixed combustor were simulated using CFD method. The combustor was separated in several regions, includingpreheating region, flame front region, flame transition region, post flame region, central recirculation region and corner recirculation region, based on the fuel-air blending characteristics, velocity field, temperature field, OH distribution, and Damkohler number. A complex CRN model was established to characterize the flow characteristics and flame structure inside the combustion chamber. The preliminary verification of CRN model was performed with experimental data of lean premixed burner. The effects of several crucial parameters, including the residence time and the flue gas recirculation ratio, on NOx emission were analyzed through sensitivity and uncertainty analysis. The CFD-CRN hybrid method can quickly and effectively predict the NOx production of the lean premixed combustor under high equivalence conditions. The residence time disturbances of preheating zone and the flame front zone have more significant effects on emissions and stability of NOx prediction of CRN models under the same disturbance intensity. The flue gas recirculation ratio of CFD-CRN method should be accurately calculated under high adiabatic flame temperature, which has significant effect on NOx formation.