OCR Text |
Show either too little or too much subtraction showed large residuals. Note also that for the ease where no subtraction was made, the residual value follows a somewhat Gaussian curve with a peak near 2200 em-', which corresponds well with the measured Stokes laser profile. The individual coal spectra were fit by subtracting the appropriate fraction of the Gaussian dye profile from the data and fitting the resulting spectra. The magnitude of the fraction of dye profile subtracted from each spectrum was proportional to the peak counts and amount of nonresonant enhancement. A ·value· for the amount of nonresonant addition was caJculated by averaging the signal over 100 pixels of the detector array (ca 2181-2233 cm-') and expressing this average as a percentage of the nitrogen peak height. The value obtained was called ·percent background" and was conveniently used to relate the amount of nonresonant enhancement one spectrum had relative to another. RESULTS AND DISCUSSION Average Spectra The averaged spectra from both the gas and coal-seeded gas flames were fit for temperature as previously described. The gas flame spectra were fit using the unmodified Sandia fitting code, and the coal-seeded gas flame spectra were fit with both the unmodified code and the modified code which accounted for nonresonant enhancement. When spectra that contained breakdown were fit for temperature without regard to percent breakdown, the fit temperatures were artificially high. This was caused by hot band increases in height relative to the cold band with the incoherent addition. The axial temperature profiles of the natural gas and natural gas-coal flames are shown in Figure 9. Data points shown for the coal flame are the average of four measurements, each of which are 100 laserpulse averages. The uncorrected coal flame temperatures correspond to fitted temperatures with no background subtracted, while the corrected temperatures are those found through properly fitting the background. It is interesting to note that although there is a large variation in the uncorrected temperature fits at a given axial location, the corrected fits all give nearly the same temperature. Single Shot Spectra In addition to the averaged measurements, 264 single-shot measurements were taken at the 1.3 cm position in the coal-seeded flame to determine the potential for future instantaneous studies in particleladen environments. Each of the single-shot spectra was classified as to background level, and the spectra in each group added to form an average spectra at that background level. These spectra were then fit both without and with background subtraction, and gave similar temperatures (ca 1840K) when fit with the proper background subtraction. From these results it became apparent that the background level in each spectrum was independent of the actual flame temperature. In other words, a consistent temperature can be obtained for a given coal-seeded gas flame independent of the background level in 13 |