| OCR Text |
Show burner size, while the contribution from the flame-sheet region is independent of burner Slze. This scaling can be readily seen from §4, since under these condi tions EI NOXnear-burner "J D~/Qo f"V Q~/2 EI NOx flame-sheet f"V D6 f1f/Qo f"V Q8 EI NOx furnace f"V D~/Qo f"V Q~/2 EI NOXstaging-iet f"V D~/Qo f"V Q~/2 As a consequence, at the 30k VV scale, the near- burner region and the flame-sheet region make nearly equal contributions to the total NOx emissions, while the contribution from the furnace is entirely negligible. With burner size increasing as the 12 MW scale is approached, the contributions from the flame-sheet and furnace regions become comparable, but both are small in comparison with the NOx production rate in the near-burner region. There is thus a transition indicated by the model in the physical mechanisms that dominate NOx emissions over the range of scales from 30 kW to 12 MW. In the top right panel in Fig. 11, the solid circles give the model predictions for the total NOx emissions, obtained by summing over all the NOx sources in the top left panel. The solid line gives the scaling obtained from the near-burner region alone, which appears to dominate the NOx emissions under these conditions in the larger burners. The open circles give the measured emission levels at full load conditions from Fig. 3. Note that, over the range of scales from 30 k vV to 1.4 MW, the model predictions obtained under the assumption of perfect similarity are in fair agreement with the measurements. At the 4 MW and 12 MW scales, however, there are significant deviations from the scaling predictions obtained when perfect similarity is assumed. It will be seen in §6 that when departures from perfect similarity evident in the in-flame data are accounted for, the scaling model gives the results shown by the solid squares, which can be seen to predict the effects observed at the larger scales. 5.2 Burner Perfornlance Scaling Figure 12 compares the predictions obtained from the scaling model under the assumption of perfect similarity (solid symbols) with the measured N Ox emissions (open symbols) for burner performance with turndown at all five primary burner scales. Also shown are results for the 300 kvV performance in the BERL tests under both hot-wall and cold-wall conditions. The top left panel shows results in terms of absolute NOx emissions, and the top right panel presents the same results relative to the baseline for each scale. The model results correctly capture most of the trends seen in the data. This can be seen in the bottom left panel, which shows the predicted and measured NOx emissions for each point. Note that the correlation achieved is 98%. Figure 13 compares the scaling model predictions, again under the assumption of perfect similarity (solid symbols), with the measured NOx emissions (open symbols) over the entire range of air dilution levels. As before, the top left panel shows results in terms of absolute NOx emissions, with the top right panel giving results relative to the baseline for each scale. The model appears to capture many of the trends in the data, though there are clear differences. at. the 12 -NlyV scale. Th~ correlation in the bottom left panel 15 |
