| OCR Text |
Show .1.6.13 values as used for Fig. 7, except that T Q H *as reduced by a factor 4, while CN03c>eq was increased by the same factor. For various reasons, the experimental results cited do not lend themselves to a more direct comparison with the present theoretical results. V. CONCLUDING REMARKS. The models discussed in Sees. II and IV both appear to be able to account for experimental data on the reduction of NO by NH3 in oxygen-rich postflame gases. However, the assumption COH3 = constant made in Sec. II does not take account of the processes determining the OH-concentration. This is remedied at least in principle by the model of Sec. IV, in which COH3 is allowed to vary with time. This is equivalent to allowing Fenimore's critical NO-concentration [N03c to vary with time. For a given set of experimental conditions, it is possible to find an average value of [N03c appropriate for use in the model of Sec. II (cf. Eqs. 16-19). The model of Sec. II has the attraction of being simple, and may be useful for making order of magnitude estimates. The model of Sec. IV is more involved computationally. However, it is self-consistent and is believed to describe the essential features of the reduction process correctly. Both models allow extracting a value of k^k3/k2 from experimental data; but neither model allows an unambiguous determination of k3 and k2/k^ separately. Furthermore, the model of Sec. IV does not allow an unambiguous determination of oc The results obtained can be used to predict the reduction of NO by NH3 under conditions prevailing in industrial gas turbine combustors. A typical combustor pressure is 10 atm. The final concentration of NO should be 75 ppm at most; values about 4 times smaller than that are desirable. Furthermore, it is desirable that practically no NH3 be left at the end of the reduction process. This requires the value of T = t/tc introduced in Sec. II to be of order 3 or more. Equation (20) then yields a minimum residence time t of 18 ms; for [N03c = 20 ppm, t 2- 66 ms. The actual time available is only 1 or 2 ms. The NO reductions that can be obtained in such a short time are insignificant. The situation can be illustrated by assuming that k2/k^, k3, T Q H and Tjjk4 are given by the values used to generate Fig. 7. Fig. 12 shows the corresponding NO-concentration as a function of time for five different temperatures, for the case that [N03o • 100 ppm, CNH330 = 200 ppm. It can be seen that the lowest NO-concentration that can be obtained in 10 ms is about 55 ppm. The optimum temperature in this case is about 1250 K (1790 °F). In 40 ms, the NO-concentration can be reduced to 20 ppm by maintaining the temperature |
