OCR Text |
Show than 8 mesh, only two dimensional measurements were made. The third dimension was then inferred using the average aspect ratio found for the larger particles. Equivalent spherical diameters were then estimated as before. In general, particles were randomly selected from screened batches, although those which were mostly shale or pyrite were discarded. Particles sized 3 x 4 mesh (^0.64 cm equivalent diameter) were burned one at a time. Particles sized 6 x 4 mesh (MD.28 cm equivalent diameter) were burned in groups of three. Particles sized 9 x 10 mesh (^0.21 cm equivalent diameter) were burned in groups of five. Smaller particles were burned in groups of 20 or more. It is important to note that, except for two series of runs, all of these combustion experiments were run starting with raw coal. Thus the coal particles were allowed to devolatilize and become char at the particular conditions being studied. This is in contrast to many kinetics studies in the literature where the char has been initially prepared at one set of conditions. We decided to run most of the experiments with raw coal because we felt that this would be more representative of the behavior in a real combustor operated at different conditions. In other words, if the combustion conditions affect the properties of the char which results from devolatilization (e.g., the porosity, surface reactivity, etc.) then any practical kinetics expression must include these effects. As is discussed later, the combustor conditions were indeed found to affect the properties of the devolatilized char. Chemical Analysis of Kentucky No. 9 Coal A typical analysis of the coal used in these experiments is given in Table 1. In all cases, the coal was air-dried at room temperature before use. In addition to the average values, typical ranges are given for some of the components based on previous ORNL experience with Kentucky No. 9. Significant variations in ash content are not unusual. This variation was even seen to occur from particle to particle in these experiments. Thus it is important to use sample averaging to reduce the confusion resulting from this variation. 15 |