| OCR Text |
Show detectors. Data of in-flame profiles are reported on a dry basis, while integral exhaust concentrations are corrected to an inlet molar flow rate basis, in order to account for dilution effects caused by changes in number of moles during reaction. Two measurement phenomena are worth noting. First, the presence of C2H2 gave a false (positive) reading of HCN that was measured by a selective ion electrode. C2H2 and HCN are usually present simultaneously during combustion of most fossil-fuels, especially under fuel-rich conditions. Fortunately, this interference from C2H2 can be eliminated by sparging the sample solution containing HCN, with dry nitrogen. The results reported here were so obtained, and are believed to be interference free. Second, there was a significant amount of N02 (up to 80% of the total NOj present in samples containing 1850-3200 ppmv NOx and appreciable O2• This phenomenon was not predicted by the flame model, and was verified to be an artifact caused in the sample line, which had a residence time of 2 to 3 minutes. Here, NOx values are reported, to account for the total amount of NO and N02 present in the system. Temperature was measured using uncoated PtiPt-13%Rh thermocouple of 0.000127 m (0.005") diameter, with thermocouple wires extended horizontally to lie along isotherms. Customary thermocouple radiation corrections were made. RESULTS and DISCUSSION Exhaust Measurements Experimental results of exhaust measurements are presented frrst, since they demonstrate the net destruction of NO in overall fuel lean diffusion flames, and this was the motivation behind this work. Both cup mixed average exhaust concentrations (Fig 5) and corresponding percentage of NOx reduced by reburning (Fig 6) are presented as a function of primary NO content entering the system. The difference between data points and the zero destruction line (which lies at 45 0 , because of the dilution correction made to the exhaust ppm) defines the amount of NO destroyed during the combustion process. Reduction of NO is more effective at higher inlet NO contents, which is to be expected. However, not expected was the observation that NO destruction of over 90% could be achieved in Configuration B, where NO (albeit undiluted at the bottom burner) was introduced along the center streamline. In all these experiments (A through E), HCN in the exhaust was less than 37 ppmv, and NH3 was below 1 ppmv. A more quantitative test to explore effects of how inlet NO contacted the reaction zone, can be seen by comparing Configurations A through C. Results from these three configurations suggested that the larger fraction of injected NO that contacted the flat diffusion flame, the more was destroyed, up to levels exceeding 90%, as in Configuration B. However, this argument was specious, as described below. The baseline diffusion flame, realized in Configuration A, in fact includes two regimes, or two types of flame, namely, the inner, stretched flat portion, which follows a I-D model, and, beyond that, curling upward beyond the upper burner edge, the sheared co-flowing portion, |
