Methods toward improved simulations for the oxides of nitrogen in pulverized-coal furnaces.

摘要

The controlling mechanisms governing the amount of unburned carbon in fly ash from practical pulverized coal combustion systems have been studied by incorporating elements of an advanced carbon burnout kinetic model in a computational fluid dynamics based, cfd, combustion simulator. Moreover, varying compositional effects for the products of coal combustion during the devolatilization and char oxidation regime are explored through the addition of three new coal off-gas progress variables. Turbulent transport of these new mixture fraction progress variables is coupled to a prescribed multivariate joint β-probability density function, PDF, subgrid mixing model.; Experimental data sets from the large scale pulverized-coal axial fired cylindrical combustor at the Imperial College of Science and Technology, the International Flame Research Foundation's long pulverized coal flame, and the University of Utah bench scale test facility are presented to evaluate mechanistic variations in the carbon burnout process, the effect of the carbon burnout process on nitrogen oxide emissions, and the sensitivity of state space variables and NO to a varying coal off-gas composition.; The cfd based simulation results indicate that in order to accurately predict loss on ignition, LOI, the char oxidation model must contain mechanisms including effects such as thermal annealing, a developing ash film resistance, and a diminishing surface area for heterogeneous reaction. Simulated nitric oxide concentrations were also found to be highly sensitive to the choice of char oxidation submodel used.; The sensitivity of state space variables to the prescribed shape of the PDF (clipped Gaussian vs. joint β) is demonstrated. Integration robustness and accuracy issues for the joint β-PDF are presented. Mixing model results indicate that although state space variables, such as temperature and oxygen, are rather insensitive to a variable coal off-gas composition, NO predictions are not. The new generation joint β-PDF mixing model allows, for the first time, proper mixing effects in the calculation of time averaged reaction rates as rigorously required within the species transports equations of NO and HCN.