OCR Text |
Show D. Marlow and T. S. Norton greatly with varying equivalence ratio, flame residence time, temperature, and hydrocarbon concentration in the fuel. Since NH3 oxidation is sensitive to the kinetics and the kinetics is in tum sensitive to the flame environment, a kinetic mechanism should be incorporated to accurately model NOx formation in turbulent flames. A simple, empirical kinetic mechanism for NH3 conversion has been included previously in a model of a turbulent, jet flame (Drake et al., 1984). The mechanism contains two global reactions to form NO and N2• Formation of HCN is not considered. The. global rates are written as the products of Arrhenius coefficients, the NH3 concentration, and the equilibrium OH concentration. The jet is modeled for a low turbulent Reynolds number and varying NH3 concentration. Predictions and measurements for the jet show similar trends with varying NH3 concentration, but the predicted NOx emissions are consistently 50% greater than measured. Several assumptions made by Drake et ale probably limit both the accuracy and range of applicability of their empirical mechanism. First, neglect of HCN chemistry should cause inaccuracies for all but lean, premixed combustion, since HCN is frequently important as both an intermediate and a product of NH3 conversion for fuel-rich conditions. Hydrogen cyanide is especially important as a product of rich-stage combustion in a two-stage reactor since the HCN created in the first stage is converted to NO in the second stage. Second, though the equilibrium OH concentration has been applied in the rate calculations, we cannot expect to be able to correlate NH3 conversion with· an equilibrium OH concentration over the entire flame. Instantaneous data from laser-based experiments (Barlow and Carter, 1994) indicate that the region of a jet flame near its tip is nearly equilibrated at the state defined by the instantaneous mixture fraction. However, near the nozzle flamelet-like behaviour occurs, resulting in OH concentrations far from equilibrium. More detailed descriptions of both the nitrogen and hydrocarbon oxidation kinetics are needed, especially to accurately model the influence of nonequilibrium radical concentrations on NOx conversion. Though rates and minor pathways may still be controversial, the detailed kinetic mechanism for low-heating-value gas combustion is well understood at atmospheric pressure. However, due to cOmputational limitations the full mechanism cannot be used for industrially relevant flames. Peters (1991) has developed a method for reducing the detailed mechanism. First, reactions are eliminated from the full, detailed mechanism to create the minimum possible reaction set (called a skeletal mechanism) capable of accurately modeling the system of interest. Then partiatequilibrium and steady-state assumptions are applied to create algebraic expressions for shortlived intermediates, and global reactions are formulated for the remaining non-steady state species. Because reduced mechanisms are simplifications of detailed mechanisms, they provide computationally tractable models that are more closely representative of the full chemistry than empirical, global rates. Several researchers are actively developing reduced mechanisms for NOx formation. Chung and Lee (1993) have modeled laminar, counterflow diffusion flames of natural gas. Only the 2 |