OCR Text |
Show DISCUSSION The results make intuitive sense in that as flame strain increases with excess air or swirl intensity, the reaction becomes less luminous, i.e. reduced radical emissions and radical population. Indeed, this relationship has been shown to follow an exponential increase with heat release rate and reaction temperature (Samaniego, et al., 1995) as well as equivalence ratio (Dandy and Vosen, 1992). Due to this equivalence ratio and temperature relationship, it is not surprising that N O x correlates so well with O H and C 0 2 radical emissions at the burner throat. Using these light emissions as indicators could provide very fast N O x information for active control. It was expected that a fluctuating component of some radical species would provide an indication of global C O emission. This was based on the hypothesis that poor mixing or quenching would cause intermittent heat release (radical formation). This was not borne out by the measurements. Upon further reflection, however, it was recognized that (1) the measurements represent a spatial average of species and (2) the chemiluminescence signals decreases with increasing strain (i.e., conditions of potential increases in exhaust C O ) . In order to adjust for these considerations, the fluctuating components were normalized by the average values to provide a percent fluctuation level. These plots are provided in Figure 8. a) Co-swirl Injector 0.6 0.7 0.8 0.9 Swirl (S') 0.6 0.7 0.8 0.9 Swirl (S1) b) Counter-swirl Injector Figure 8: O H and C 0 2 % Fluctuations The co-swirl injector exhibits high levels over most of the operating range. This may be due to the overall poor mixing and unsteady reaction present for this injector. The counter-swirl injector displays the best correlation to the C O emissions along the high swirl stability limit with the O H 9 c fluctuation measurements. It is interesting to note that the C02%fluctuations, unlike all of the other cases, does not follow the same trend as O H . |