OCR Text |
Show the furnace to be either completely clear of seed or filled with seed to uniform concentration. By delaying the time when the image was recorded from the beginning or end of the seed pulse, either light-on-dark or dark-on-light images of the mixing region could be obtained. In general, the images in which the incoming fluid was seeded and the furnace fluid was not (light-on-dark) were of higher quality and are therefore presented here. Note that although the flow of TiC14 was started and stopped impulsively just prior to each measurement, we do not believe that this caused a significant perturbation to the flow structure since its mass flowrate was small in comparison with the air. Single component, dual-scattering laser Doppler velocimetry was used to establish the time-mean and RMS velocity field of the flow near the inlet. Table n lists the operating characteristics of this system and a complete description may be found in Edwards (1989). RESULTS We begin by looking at the time-mean structure of the swirling jet as determined by point LOV measurements. Figure 3 shows the mean and mean +/- RMS values of the axial component of velocity at the furnace entrance. The mean velocity shows an annular jet proceeding in the positive z direction with recirculation along the centerline. The RMS velocity fluctuations are a substantial fraction of the mean velocity throughout the jet and recirculation regions. Figures 4 and 5 show the same information for the tangential and radial velocity components. From these figures one might conclude that the flow is fairly symmetric and well behaved, but with a reasonably large fluctuating component. In fact, the structure of the flow is quite different than this. Figures 6 through 9 show four images of the elastic scattering off of Ti02 seeded into the incoming fluid. Each was taken at the same conditions (650 ms delay from start of seeding) and images the flow through a diametral plane of the furnace in the streamwise direction. The field of view covered by the images is 165 mm in radius and extends 225 mm from the inlet plane. The four realizations show that the instantaneous structure of the flow is neither very symmetric nor well behaved. In reality, the flow is made up of structures with length scales of the order of the throat diameter (100 nun) which form within one throat diameter of the entrance plane. Note that as the fluid first enters the furnace, structures resembling those produced by the Kelvin-Helmholtz instability in a two-dimensional shear layer begin to form at the outer edge of the throat. However, unlike the simple shear layer these structures do not continue to grow monotonically with distance but are soon rolled into an eddy of significantly larger scale. These larger eddies are convected downstream along paths with increasing radii and appear to alternate in location from the left to right side of the flow. Images taken in the cross-stream plane indicate that these structures may be connected in a helical fashion. Figure 10 shows an image taken in the plane one-half a throat diameter from the furnace entrance (z = dI2). The structures in the cross-stream plane were imaged onto fum from above at a 45° angle. As a result, a circle appears as an ellipse in the images. Two features of this figure are significant. The first is that there are structures with a length scale of order 10 mm which form as a result of shear in the azimuthal direction. The second is that the ellipse, which represents the circular cross section of the flow, appears to have a tail which proceeds -4- |