OCR Text |
Show hours. The input to the program consists of: 1) geometry data; 2) gas transport, thermodynamic and chemical properties; 3) reaction chemistry information; 4) radiation parameters; and 5) program control variables (iterations, relaxation, etc.). Model Results A standard North American 4425-SA nozzle mix, air-fuel burner (Figure 1) was fired in the Air Products combustion lab furnace (Figure 2). The parameter ranges tested were as follo~~s : 21 2S% oxygen levels in the air stream 1.5 - 2.5 MMBtu/hr firing rates 1.S - 2.5 stoichiometric ratios The furnace was fired at each set of conditions until steady state was achieved (one half hour minimum in a warmed up furnace). A positive furnace pressure was maintained to exclude air infil tration in order to isolate the effect of the burner on NOx formation. The model is currently being used to predict the NOx produced by this burner in the Air Products combustion lab furnace. Since GRREK is designed for 2 dimensional cylindrical coordinates, the rectangular furnace is approximated by a cyl inder of equal cross-sectional area. A plot of the grid is shown in Figure 3. The grid consists of 36 nodes in the axial direction and 31 nodes in the radial direction. The furnace is 17.7' long and has an equivalent radius of 4.2'. The exhaust flue is 2' in diameter. Although GRREK is designed to handle any number of reactions, only the prompt NOx and thermal NOx reactions are being considered since fuel NOx is insignificant in natural gas combustion processes. The test cases being modeled are shown in Table 2. Case 1 is the base case with the furnace wall temperature at 20000 F with the full set of prompt and thermal NOx reactions. For case 2, the wall temperature was lowered to lOOooF to see the effect on NOx. The flue NO is lowered because the increased radiation from the flame to the ~'alls lowers the overall flame temperature which reduces the thermal NOx. Because the wall is some distance from the flame, the effect is not very significant. Case 3 is the same as case 1 except that the prompt NOx reactions are removed so that their relative importance can be assessed. Table 3 shows the reduced set of reactions suggested by Pratt and Wormeck9 • The flue NO is reduced, but not significantly. Therefore prompt NOx does not appear to be important in this particular combustion system. -5- |