| OCR Text |
Show furnace designers understood the importance, to good mixing and to control of combustion pr01 gress, of the input feedstream momentum. My own status was one example. In a graduate class in furnace design I used to emphasize the importance of mixing and its relation to input momentum, adding that complete combustion with 15 to 20% excess air was a reasonable goal. Decades later German steam boiler practice showed that, with more input momentum, 5% excess air was a~ quate and 1% was feasible. Today, of course, the Ijmuiden reports identify not only momentum ~nitude but swirl number as well, and classify flame types as type 3, high axial momentum with little or no swirl, a swirl number of 0 to 0.5; type 2, with very high swirl number to produce internal axial recirculation; type 1, a mixture of types 3 and 2. And to me the most exciting prospect coming out of the International Flame Foundation1s work is the next-3-year study labeled NFA,- near-field aerodynamics of swirl burners. Flame placement of course has a profound effect on heat-flux distribution in the combustion chamber, and heat transfer is the reason be build furnaces. I come now to the only part of my comments that conatins something that may be useful to you. The papers of this two-day symposium include 9 related to oxygen enrichment, and I thought it would be worthwhile to generalize about the effect of air enrichment with oxygen on the nonluminous gas emissivity of the resultant combustion products. Methane was taken as a typical fuel, burned with one percent excess oxidant (quite feasible when the added oxygen is large, more difficult if none has been added).The variables affecting emissivity EG are the furnace size, measured by the average mean radiant beam length L eft], the gas temperature T[OR], and the added oxygen, best presented as the fraction of the total incoming oxygen that does not enter as air. Putting three independent parameters into the picture can produce a mess. Because the pr~ duct (emi ss i vi ty p< tempera ture) ,EGT, is more nearly independent of T than EG alone, EGT curves at three widely separated T1s(1800, 2700, 3600R) record the full temperature effect. Four different furnace sizes, for path lengths L of 5,10, 15,20 ft., are given. With EGT as ordinate vs. the fraction of 02 not coming from ai~ 12 curves tell the whole story (see Figure which follows). If one focuses on EGT at a mean T of 2700R--the dashed curves with square data points --the effects of furnace size and oxygen addition stand out. Example: Increasing non-air 02 from o to 0.5 to 1.0 when L = 10 feet changes EG at 2700R from~912 to 1097 to 1542)/2700, or from 0.34 to 0.41 to 0.57. At 3150R, halfway from 2700 to 3600, emissivity varies from (935 to 1140 to 1670)/3150, or from .30 to .36 to .53. Note that the numerator did not change much and can be read by interpolation with good accuracy. Note also that as temperature rises, wi~ Land 02 constant, emissivity always goes down. 5 Although adiabatic flame temperatures can be very high when the oxygen is high, one must remember that mixing with combustion products must be intense in a well designed system, that the gases are losing heat as they radiate, and that mean temperatures are therefore far below adiabatic flame temperature. The basic relations used for computing tt-e curves are those appearing on the plot; they were obtained by fittin9 functions to the numerical recommendations coming out of Dr.I.Faragls thesis study, as modified by Sarofim and me. |
