OCR Text |
Show As shown on Figure 3, a second jet of cooling air is directed onto the bottom of the target. This cooling air exits from a flow annulus concentric to a sapphire rod. The sapphire rod is coupled by optical fibers to a two color pyrometer which serves to measure the temperature of the bottom of the target. By adjusting the flow of cooling air, the temperature of the target could be reduced. All target temperatl'res reported herein correspond to the steady state temperature of the clean target, measured as described above, without the coal feed turned on. The coal feed had a minor effect on the surface temperature, typically raising the surface temperature by SOK. The coal feed was deliberately run dilute resulting in an ash flux aproximately ten times smaller than in an operating turbine. The low feed rate allowed deposition testing to be stretched over longer times, without a significant change in reactor temperature, thereby improving the data consistency. Moderate variations to the coal feed rate were tested and showed that changes in the coal feed rate had little effect on the measured sticking coefficient. Experiments proceed by inserting a clean target of known weight into the jet and allowing a deposit to build up for (typically) 10 minutes. The ash arriving at the target during this same time interval is known from filter samples collected earlier. The sticking coefficient is the ratio of the increased target weight to the weight of ash which arrives at the target. Filter samples were obtained by inserting a 47 mm filter in the same location as the target, with a vacuum pump used to draw gases through the filter. In this manner, all the ash exiting the reactor nozzle was captured in the filter, and could be weighed over the same ten minute test interval. Some provlslon was needed to account for unburned carbon appearing in the (cold) filter, which was otherwise burned out of the (hot) deposit. This correction was performed by a routine carbon analysis of the filter sample, with appropriate correction of the ash arrival rate used to calculate the sticking fraction. Results Figures 4 and 5 show the measured sticking coefficient at two different reactor temperatures, 1373K and 1573K respectively. At the lower reactor temperature (Figure 4), the sticking coefficient is independent of the target temperature. At the higher reactor temperature of 1573K (Figure 5), the sticking coefficient is seen to decrease by an order of magnitude as the surface temperature drops from 1230K to 1070K. This reduction in sticking coefficient with surface temperature may be the result of a number of physical processes occurring at the target. First, the lower surface temperature could freeze molten ash particles as they traverse the boundary layer above the cooled target. Second, any chemical reactions producing molten mixtures in the deposit would be substantially slowed as the temperature of the target/deposit is reduced. Vhile the mechanisms which contribute to the behavior in Figure 5 are easily understood, a comparison of data in Figures 4 and 5 reveals some unexpected results. Specifically, it should be noted that the sticking coefficient at the higher reactor temperature (1573K, Figure 5) is consistently lower than at the low reactor temperature (1373K, Figure 4). This effect is most noticeably observed by comparing results in Figure 4 and 5 at the target temperature 1050K. The sticking coefficient differs by two orders of magnitude at these conditions. The latter result is somewhat surprising since it was initially thought that the cooler reactor temperature would have a lower sticking coefficient |