OCR Text |
Show steady state, however, required very long waiting times, around 5 to 10 h, longer than could be afforded, so data collection was generally started when refractory temperature levels were within about 100 °C of the steady-state values, usually 1 to 2 h after starting. The small but still not quite negligible transient effects (heating of the refractory walls) were modelled and taken into account in the energy balances. The satisfactory accuracy of these estimates was experimentally demonstrated in a number of cases. The energy balances have no novelty and are not dwelt upon in the present account. Their main significance here is that they provide the estimates of the mean exhaust gas temperature, Te, at combustion chamber exit. In these calculations, the exhaust gases were assumed to be in chemical equilibrium. The air preheat was obtained, as w e have noted, by recuperation. Because of thermal capacity in the exhaust train, from the combustion chamber exit up to and through the recuperators, and then the additional capacitance in the combustion airline thereafter, the combustion air temperature reaching the burner(s) considerably lagged the combustion-chamber refractory temperatures in the approach to steady state. This provided an opportunity, which was fully exploited, for studying the effect of air preheat on N O * emission "on the run", while the air preheat gradually ascended, the furnace gases (with little thermal capacitance and only a short residence time) being always in quasi-steady state in respect to their boundary conditions in the combustion chamber. The amount of data so collected on the effects of air preheat greatly exceeds that which could reasonably have been obtained, within the constraints of budget and time, by restricting data-taking to periods of full steady-state. 7.2. The NOx emissions 7.2.1. The Series 1-XBM2-Tsingle-burner, tunnel-chamber trials In these trials, the water-cooled floor panels were completely covered with insulating blanket. With the sink thus heavily screened, the furnace gases and the refractory surfaces (roof and walls) reached elevated temperatures, the refractory showing maximum (hot spot) temperatures as high as the operating limit, 1350 °C. Exhaust gas temperatures (leaving the combustion chamber) were as high as 2000 °C. The largest set of data for a given burner configuration was obtained, by design, in the two trials with D\ = 6.35 m m , Di = 19.05 m m and 6\ = 30°. Figure 5 shows the N O x level in the exhaust gases as a function of the O2 content and the air preheat with the firing rate at the practical maximum, 289-368 k W (the maximum was determined by the air supply pressure and thus varied somewhat with the air preheat and excess air level), and with the pilot burner on. The negative O2 levels are, of course, not real but, as explained in § 4, represent negative excess air as indicated by the resultant C O , with every two C O molecules counting as one molecule of "negative O2". The relations between the exhaust-gas contents of O2 and C O and the excess air level are also given in § 4. Roughly, e = 5 X0i at s > 0. At X0 = 0.03 (or 3 % ) , s =0.150 exactly (or 15%). The obvious trends seen in Fig. 5 are for the N O * to slowly rise with increasing positive O2 (positive excess air), and to quickly decrease with increasing "negative O 2 " (negative excess air). The expected increase of N O * with increasing air preheat is substantial. Additional data were obtained at firing rates down to 77 k W , or 1/5 the maximum, and with the pilot burner off as well 8 |