| OCR Text |
Show exhibits higher relative intensity at higher frequencies. In our opinion, this indicates much higher activity of small eddies and, therefore, higher degree of small-scale turbulence and mixing rate in 10w-NOx burners. 7. Our main conclusion was that temporal frequency flame spectra provide sufficient basis for developing criteria for optimizing individual burners. These criteria can be based on the . quantitative analysis of statistical parameters characterizing the pattern and shape of the spectrum, as it was suggested in our previous publications (Ref. 5,6). We also observed that flame spectra often contain specific fine features at certain frequences, which may also be utilized to form the above optimization criteria. 8. Our results have confirmed the feasibility of the proposed concept. The next important step was to organize testing on single-burner well instrumented test facilities to investigate correlations with burner parameters, such as burner stoichiometry and NOx. Testing at single-burner combustion test facilities. The main purpose of these tests was to determine whether we could prove that the statistical parameters we extract from the temporal flame spectra can be positively correlated with burner parameters, such as stoichiometric ratio, NOx, etc. The only reliable way to study such correlations was via measurements taken on well-instrumented pilot-scale single-burner facilities. Fortunately, we received several opportunities for such testing. Data were obtained from three well equipped coal-fired combustion test facilites of different sizes: 2.0, 5.0 and 100 MBtulhr. All three facilities were equipped with different 10w-NOx burners of latest design, allowing a wide range of air distribution between primary, secondary and tertiary (overfire) air ports, and all were equipped with instrumentation for measuring all coal and air flows, and with necessary flue gas analyzers. Our tests allowed us to make a number of interesting observations. We observed the transformation of the measured flame traces with excess air and established some specific features which may have some significance. Our main conclusion from the test data on coal-fired facilities is positive and optimistic: we confirm that by properly processing the temporal frequency flame data, we are able to generate on-line information characterizing burner flame stoichiometry and NOx. This conclusion is illustrated below by several examples for coal-fired 10w-NOx burners. Figure 6 illustrates the correlation between our calculated burner flame quality parameter Q and NOx with variations of burner inner swirl, for a pilot scale coal-fired 10w-NOx burner 5 f\.1Btulhr. Variation of the inner swirl significantly affects burner NOx, and the response of our flame quality parameter is well correlated with the direct measurements of a NOx analyzer. Other burner flame quality parameters Qi are well correlated with burner stoichiometry. Figure 7 illustrates the correlation between our calculated burner flame quality parameter Q and NOx at two different loads, for a combustor with coal-fired 10w-NOx burner 100 MBtulhr. Figure 8 illustrates the correlation with NOx during redistribution between secondary (SA) and tertiary (TA) airflows for a pilot scale low NOx burner 2.0 Mbtulhr. Initially, TA was about 64 % of the total air flow, and SA was about 17 %. During the tests, TA flow was incrementally reduced 9 |
