OCR Text |
Show 8 in Figure 4, Visser and Levinsky's curve fit (adjusted to 15% 02) is matched by the following exponential equation: NOx ppmvd (15% 02) = (1.58E+5)exp( -19630rr(K)) The results of Altemark and Knauber (1987) are for the SINOX prototype burner. The pressure range covered in these experiments was 4 to lObar; over this range the NOx showed essentially no sensitivity to pressure. The measurements, however, did show a sensitivity to inlet temperature. In Figure 4, it is shown that the change from 673 to 473K inlet temperature reduced the NOx by about 20% on average. The flame temperature is that computed by Altemark and Knauber. We adjusted the NOx data from dry, actual 02 basis to dry, 15% 02 basis. The Altemark and Knauber data indicate a lower apparent activation energy of NOx fonnation than given by the other data. The argument of the exponential curve fits to their data is about (-16320rr(K)). It is conjectured (herein) that this behavior may have been due to heat loss from the burner at the highest temperatures tested. Results of Aigner et al. (1990) for the ABB double cone burner (late-1980s model) operated at engine conditions are also shown in Figure 4. We computed the adiabatic equilibrium flame temperature assuming methane and using the fuel-air equivalence ratios given in the paper. On average, these NOx data show about the same flame temperature sensitivity as the Sattelmayer et al. (1990) data. For the data of Aigner et al. the argument of the exponential curve fits is (-25570!f(K)) on average. Two conclusions which can be stated based on the data in Figure 4 are the following: 1. For the most part, the stirred reactor data (from 1.7/1.9ms to 6.0/6.9ms nominal residence time) bracket the other atmospheric pressure data and the data for which the pressure effect appears to be weak to negligible (i.e., the data of Leonard and Stegmaier, 1993; Visser and Levinsky, 1993; and Altemark and Knauber, 1987). This suggests that the stirred reactor is a useful laboratory tool for studying and seeking an understanding of NOx formation in high-intensity, lean-premixed combustion. 2. The arguments of the exponential curve fits given above vary from (-16300rr(K)) to (-2869SrrCK)). These correspond to apparent activation energies of NOx formation of 32 to 57kcal!gmol. The average apparent activation energy of 44.5kcal/gmol is nearly identical to that of the 590/600K stirred reactor data of Figure 3, i.e., 44kcal/gmol. Some of the differences in the data of Figure 4 are very likely due to uncertainties in the flame temperature. The adiabatic equilibrium temperature is an estimate, or engineering representation, of the actual temperature affecting the NOx fonnation. Since the NOx is a strong function of the flame temperature (even in this regime), lack of precise knowledge of the temperature can affect the comparison and interpretation of the NOx data. It should also be noted that there is a distinct difference between the stirred reactor and the other burners with respect to effect of residence time. In the stirred reactor the highly non-equilibrium chemistry is forced to fill the whole reactor. Thus, the NOx scales with residence time. In the burner of Leonard and Stegmaier (1993), for example, this is not the case. Their statements imply that the residence times of 2 to lOOms had little effect on the NOx. For the laboratory burners used by Leonard and Stegmaier, it is unlikely |