OCR Text |
Show 12 the interior walls of the containment pipe. The LDV measurements made in 3 1 this test stand show similar asymmetries in gas velocity profiles. (Analysis of these spectra before they were normalized by the instrument transfer function indicated only marginal differences: temperature deviations between the two sets of spectra were seldom greater than 20 K. Fig. 7 displays a fit to data taken 10 cm from the center of the gas flow at location 2. The underestimate of the model of both the cold and hot nitrogen bands is indicative of lower and higher temperature components respectively contributing to the data. This fit is characteristic of those obtained in regions of high thermal gradients. The 0.5 cm spatial resolution in this region would indicate a spread of ** ± 75 K around the best fit temperature for the * 300 K/cm gradient. However, this and similar spectra indicate that temperature components outside this range are present. Consequently, spectra of this type are interpreted as 3 1 mixes due to both spatial and temporal variations. LDV measurements indicate increased flow instabilities near the wall which would lead to a greater temporal variation in temperature. Temperature profiles of the gas stream were also taken at the four locations shown in Fig. 1 at a combustor stoichiometry of 1.05 when coal fly ash (Illinois #6) or an 8:1 mixture of potassium seed (K~S0,) and fly ash were injected. Fly ash could be injected at a maximum rate of 7.6 x 10 kg/sec corresponding to a 100 percent ash carryover condition. The _3 seed and fly ash mixture was injected at a rate of 1.64 x 10 kg/sec to yield a 1 percent potassium mass concentration and a 25 percent ash carryover-the expected condition of the post-magnet gas stream in an MHD power plant. When either the fly ash or the seed and fly ash mixture is injected, gross interferences in the CARS spectra appeared. An example of this interference is shown in Fig. 8. In this trace the cold and first hot band of the CARS spectra of N are clearly visible on top of a peaked background. Several features in this figure are noteworthy. First, the background is peaked at the same wavelength as the center wavelength of the CARS shifted dye laser. This suggests that the signal has a coherent character. That is, an increase in the nonresonant susceptibility in the interaction region would produce a signal with the Stokes beam spectral profile at anti-Stokes frequencies. Second, the background and the resonant CARS spectrum of N2 are not coherently mixed. Specifically, the |