| OCR Text |
Show AFRC - 1996 Int. Symposium Baltimore 4. RESULTS 4.1 Flame stability September 30 - October 2 1996 Page 5 of 14 In the experiments, two "stable" situations were achieved: either the gas/air/flue-gas mixture was completely burnt, or no combustion occurred at all. Intermittent, unstable, flames were not observed. For all flames considered, the CO emission was below 20 ppm, indicating that combustion was completed. Under certain conditions, the flame was blown off several minutes after an adjustment of the flows. During that period, the temperature measured at the end of the flame tube decreased slowly. As soon as the flame was actually blown off, a rapid decrease in temperature from about 950°C to below 850°C was observed, causing an automatic shut off of the gas valve. Each experiment was initiated with a flame in the flame tube. The natural gas flow injected in the venturi was then reduced with small adjustments. Following each reduction, it took 2 to 3 minutes to stabilize the temperatures in the flame tube. After stabilization, a measurement was performed (duration 2 minutes), and the gas flow was further reduced. This procedure was repeated until the flame was blown off. Figure 2 (next page) presents the results showing the boundaries at which combustion either occurred or did not occur in the experimental set-up. The horizontal axis represents the fuel concentration in the natural gas/air/flue-gas mixture before combustion. The vertical axis represents the temperature of this mixture. The data presented are from stable flames at conditions close to those at which the flame was blown off. The experiments presented in figure 2 have been performed at constant mixture velocity of about 1.1 m/s at the flame holder, and for a range of oxygen concentrations in the mixture between 12 and approximately 19 %. (The latter value corresponds with a natural gas/air mixture without flue-gas) Figure 2 also contains a set of data with a calculated laminar burning velocity (Su) of 0.30 + 0.02 m/s. These calculations have been performed using a one-dimensional flamecode from Wamatz (1983) and were carried out using methane as fuel. As mentioned, methane has a laminar flame velocity which is very close to that of Groningen natural gas. In the calculations, presented in figure 2, the air factor was fixed at 1.2. The percentage of fuel was varied by "adding" a quantity of (stoichiometric) flue-gas to it. It may be noted that, using this value of laminar flame velocity, there is a reasonably good correlation with the experimentally obtained stability range. Inasmuch as flame stretch theories predict an Su-2 dependence for flame stability (Law, 1988), we expect blow off to occur along a contour of constant Su when using a constant exit velocity. It is interesting to note that the contour of constant laminar burning velocity is essentially linear when plotted in this fashion. |
