Prediction of NO x control by basic and advanced gas reburning using the Two-Stage Lagrangian model

Abstract

Natural gas reburning for NO x control is currently a mature technology which has been successfully demonstrated at full scale on numerous occasions. Recent reburning research has revealed that advanced reburning, i.e., injection of an N-agent into the reburning zone, can significantly improve the efficiency of NO x reduction. This paper presents results of the application of the Two-Stage Lagrangian (TSL) model of Broadwell and Lutz to the basic and advanced reburning processes in a 300-kW natural gas fired Boiler Simulator Facility (BSF). The injection of the reburning fuel and overfire air is modeled as independent deflected jets and the TSL model is applied while each jet completely mixes with the main flue gas stream of the boiler. The entrainment rate to the jet, which is required as a model input, is derived from control volume analysis using the experimentally determined jet-trajectory. The rest of the facility is modeled as a plug flow reactor (PFR), while the droplet injection of N-agent is modeled as being distributed and thus instantaneously mixed. Calculations were conducted with a detailed chemical mechanism based on GRI-Mech-2.11 with additional reactions characterizing thermal DeNO x chemistry and modified HCCO + NO reaction rates. The TSL model allows continuous chemical reaction during the mixing while preserving some of the important characteristics of the jet. The model predicts 50 to 80% NO x removal depending on the thermal input of the reburning fuel, initial concentration of NO x , and injection temperature of the overfire air and N-agent. The comparison with the experimental data shows good agreement for relatively high reburn zone stoichiometric ratios (SR ∼ 0.99) while the removal of NO is overestimated for richer conditions (SR ∼ 0.95 or lower). The model in general follows the trend observed in the experiment and is able to quantitatively predict the NO removal in the gas reburning process. The improvement of these predictions over previous modeling efforts using a PFR and time-distributed instantaneous mixing, suggests the importance of modeling the mixing in reburn calculations.