| OCR Text |
Show AFRC - 1996 Int. Symposium Baltimore . 1000 --~0" 100 c: .2 U 10 :J ~ e ~ >< 0 Z 0. 1 ..-, Ia~ 0 ... °a otp ~ ~. .... a September 30 - October 2 1996 Page 12 of 14 o 1D _ • II ~ 4.00E-04 4.S0E -04 S. 00E -04 S.SOE-04 6. 00E -04 6 .S0E -04 7 . 00E -04 1/Tad (11K ) IX19~02 ; 20'C a16~02 ; 400 -C .14~02; 370'C o14~02; SOO 'C .12~02 ; 360'C I Figure 7: HOx production versus l/Tad. Same data as in figure 5. mechanism, and complicated by the effects of the actual temperature trajectory on the NOx formation rate. It is expected that the Fenimore contribution varies from 15 ppm under near stoichiometric conditions to less than 2 ppm at more than 20 % excess air (Mokhov and Levinsky, 1996). Given the various mechanisms and physical effects that contribute to the total production rate,it is rather remarkable that plotting the data as a function of the flame temperature gives such a little scatter. 4.3 NOx posUJrocessor The NOx postprocessors in CFD codes contain NOx source tenns in units of mole/trills, which implies that there is a relation between the NOx production and the residence time of species in the flame zone. Given the uncertainties in the modeled time-temperature histories and the (super-equilibrium) oxygen-atom concentration, the curve fit presented above can be used to build a novel NOx source term which relates the "integral" NOx production to the amount of natural gas burnt per unit of volume. This work has been carried out at the IFRF. The NOx source tenn, given by Visser and Levinsky (see relation 2), has been implemented in a CFD-package and tested against |
