| OCR Text |
Show The CHEMKIN thermodynamic and transport database were used in the model, and a detailed reaction mechanism, containing 47 species and 212 elementary chemical reactions, was extracted from the work of Miller and Bowman[12]. C2 hydrocarbons were included, while any hydrocarbon higher than C2 and certain three reactant reactions (e.g., C2H + O2 = 2CO + H) were excluded. Axial profiles of temperature, velocity, f/!, species concentrations and rate of reactions could thus be predicted, with no adjustable parameters. .The cup mixed, exhaust concentration of NOx and, therefore, the net destruction of NO were also calculated by integration of axial profiles of the product of species concentration and radial velocity. The model included heat loss to the surroundings by gas radiation, and thermal transport caused by inter-diffusion of species. It used a pseudo binary diffusion model for the multi-component mixture, instead of true mUlti-component diffusion. EXPERIMENTS PERFORMED Diffusion Flame Configurations Figure 3 shows the schematic experimental setup of a bench-scale laminar counter flow diffusion flame. The flame was supported between two identical cylindrical burners (O.1016m or 4 inches in diameter) which were coaxially mounted, vertically opposed to each other with a fixed separation (0.0227m or 0.8937 inches apart). A bronze porous disk at each burner exit removed any non-uniformity in the flow upstream. Fuel (CH. + NJ was fed through the upper burner, while oxidant (simulated flue gas, 02+C02+ NJ came from the lower. With a suitable ignition source, a self sustained, stretched, laminar diffusion flame could be established near the boundary-free stagnation plane where the two impinging streams encountered each other. A 2-D probe positioner was employed to control the movement of any necessary sampling device (namely, thermocouple and quartz probe) between two burners. This entire set-up was housed inside a closed chamber (of O.8m in diameter and O.8m in height) which had only one exhaust vent and one access sight window. For in-flame sampling, the position of the probe (and/or thermocouple) was indicated by a He-Ne laser beam, and then measured by a cathetometer (of 5. Ox 10-sm accuracy). In order to examine the hypothesis, that the fraction of injected NO that contacted the diffusion flame front was important in determining the overall amount of NO destroyed, the lower (oxidant) burner was configured in several ways, as shown on Fig 4. In Configuration A, NO was well mixed with the oxidant and uniformly distributed throughout the entire burner. In Configuration B, a small quantity of pure NO was injected along the center streamline of the oxidant directly into the flame front, using O.OO3175m (1/8") stainless steel tubing. In Configuration C, a O.0508m (2") diameter tube was coaxially arranged inside the lower burner. The oxidant composition in both coaxial tubes was identical, except for a known concentration |
