OCR Text |
Show Thermal In pu t Tl T2 Ll/Do L2/Do [02h [02h TJ [02]J (MW) (K) (K) (%) (%) (I{) (%) Primary test scales (burner/furnace similarity) 0.03 1910 na 2.4 na 8 na 1825 5 0.3 1840 na 3.2 na 3.6 na 1575 3 1.3 1820 1860 1.8 4.8 3.5 1.5 1550 4 4 1880 1870 1.8 5 0.45 0.7 1700 2.5 12 1700 1820 1 4 0.1 0.6 1625 1.5 Secondary test scales (confinement effects) BERL - hot wall 1900 1920 1.8 4 1.5 1.5 1600 1.5 BERL - cold wall 1830 1820 1.9 3 1 1 1525 3 Table 2: Parameter values used in the scaling model in §6 to correct for imperfect similarity that resulted from the constant-velocity scaling principle on which the SCALING 400 tests were designed. All other parameter values were identical to those used in §5.1 and §5.2 under the assumption of perfect similarity across all burner scales. size Do. This is presumably due to failure of the constant-velocity scaling principle on which the tests were designed, as discussed in §5.3. Thus, for example, the size of the near-burner region and the temperatures and oxygen concentrations within this region, showed departures from the simple scalings in §4. Similarly, under fuel staging conditions the in-flame data showed variations with scale in the temperature and oxygen concentrations in the staging jet region. The variations in these parameters with burner scale are summarized in Table 2. To account for these imperfections in aerodynamic and thermal similarity under constant-velocity scaling, the values in Table 2 together with the actual dimensions of each of the test furnaces were used in the scaling model, and results analogous to those in §5.1 and §5.2 were computed. The resulting predicted NOx emission scaling with burner size is shown by the solid squares in Fig. 11. As can be seen in the top right panel, when these corrections for imperfect similarity are included, the model predicts the NOx emission to peak at 4 1tI\V and then decrease slightly at the 12 MW scale, in good agreement with the measured results. Though the scaling model still slightly over-estimates the NOx emission from the larger burners, the major conclusion to be drawn is that differences between the model results in §5 (solid circles) and the measurements (open squares) indeed appear to be largely due to the imperfect similarity in the near-burner region resulting from the constant-velocity principle on which the SCALING 400 tests were designed. vVhen these similarity departures are incorporated into the model, the agreement with measured NOx emission levels is quite good. The correlation in the bottom left panel between the scaling model predictions and measured NOx levels is 91 %. 7. CONCLUSIONS The testing portion of the SCALING 400 study has provided a unique set of data offering previo1,Jsly unattainaQl~ insights into the physical mechanisms responsible for the N Ox 18 |