| OCR Text |
Show 10 temperature profiles obtained using this geometry were not reliable. Therefore, only time averaged spectra are presented in this work. These were taken at a laser pulse rate of 10 Hz while the OMA integrated the signal for 1 second. Fig. 5 displays a CARS spectrum of molecular nitrogen taken near the center of the gas stream at location 1 while the combustor was operating at a theoretical air/fuel ratio of 1.05. This is compared with a computer-generated three-band spectrum calculated for a best fit temperature of 2480 K. The f/S factor was fixed to a value corresponding to a nitrogen mole fraction of 0.717 and a nonresonant suceptibility (XNR) °f 10.86 x 10 8 cm3/erg. A weighted average was used to calculate xNR based on predicted mole fractions and nonresonant susceptibilities measured by 2 9 3 0 21 Rado and Lundeen et al. as modified by Farrow and Rahn. Although these fits are acceptable, somewhat better results are obtained when the f/S factor is permitted to float. This procedure, however, often led to - 1 7 3 unphysical results (mole fractions > 1 or XNR < 10 cm /erg) and was judged unreliable. Accordingly, the f/S factor was fixed to obtain the temperatures given below. Fig. 6 shows radial temperature profiles at the four locations indicated in Fig. 1. These were taken in a clean gas stream (no seed or fly ash injection) when the combustor was operating at a stoichiometric ratio of 1.05. Temperatures between 350 K and 2500 K were observed. The qualitative features of these data are reasonable. Location 2 has a slightly more rounded temperature profile than does location 3 since its thermal losses are higher at the higher temperature. The temperatures at location 1 have more uniform values across the gas stream than those at location 2 because the refractory lining is four times thicker at location 1 than at location 2. Near the center of the gas stream, no statistically significant differences were found between data taken when the optics ports were unpurged and purged. (Analysis of twenty-five spectra taken near the center of the gas stream at location 2 yielded a standard deviation of 50 K). A significantly lower gas temperature was observed near the edge of the gas stream when the purge was applied. Care was exercised to ensure that the measurements were made symmetrically about the center of the gas flow. Therefore, the left-right asymmetry seen near the edges of these temperature profiles may be due to uneven slag accumulation on |
