OCR Text |
Show counting accuracy of the PCSV instrument for the largest particles. For example, a 20% error in the mass measurement corresponds to a 6% error in size measurement, assuming that all the error is attributable to the size. In fact, errors should also be attributed to statistical counting of the few large particles, non-uniformity of particle flux across the slot, and other random variations in the particle feedrate. All cold flow distribution measurements shown in this paper have been verified by independent vacuum sample measurements, showing agreement similar to that in Table II. The microscope sample measurements discussed above serve as yet another independent check of the sizing accuracy for slurry particles. Comparative Measurements for Four Slurries Results for the other slurries show similar distribution shapes, but significant differences in both total number density and evolution with increasing residence time. Figure 15 shows the number distribution results for slurry #2 as a function of reactor position, where the primary difference for #2 and #1 (Figure 9) is a factor of ten lower ash content for slurry #2 produced by chemical beneficiation. The primary effect of the ash difference is to give a much larger change in submicron particle density (2.5-3 decades) with residence time for slurry #1 (low ash) as compared to a 1-1.5 decade change for slurry #1 (high ash). Thus, there is less consumption for the submicron particles in slurry #1 as compared to slurry #2. This behavior is probably due to the higher percentage of mineral matter content which has been known to concentrate in the smaller size particles (Smith, 1980). Again, particle fragmentation is negligible for slurry #2 which also shows significant particle swelling with reaction time. Results for finely ground coal/water slurries (#3,#4,#5) are shown in Figures 12, 16, and 17. These distributions are in general similar and somewhat lower in total number density than slurries #1 and #2 which are the coarse grind slurries. The finely ground slurries show some size reduction in the submicron range, though less than that of slurry #2. For the high ash content, fine grind slurry (#3), there appears to be little change in the number frequency as a function of residence time in the size range of 3-20 microns. Indeed, there is a significant increase in the number density around 15 microns at 1.8 milliseconds. These results may indicate that some fragmentation of the larger particles is occurring or that there are relatively large inert particles in this slurry. Even if the entire effect is attributed to fragmentation, it does not appear to play a major role in size reduction of the largest particles compared to the other slurries. The low ash, fine grind slurry (#4) shows an initial large production of submicron particles followed by a rapid decrease in their population. At the large particle size end this is coupled with significant agglomeration, followed by swelling and rapid volume loss of the largest particles. Slurry #5 has similar properties as #4, showing significant particle swelling, but a lower initial submicron population. These results may be attributable to the qualitatively higher viscosity of slurry #5. Experiments at greater dilution (with known viscosity) or at higher atomization (increased shear) rates are required to confirm this inference. |