| OCR Text |
Show 5 Iron Evolution During Coal/ Char Combustion Figure 2 illustrates the evolution of iron from the coal/char as determined by these methods for the larger particle size (106-125 J1-m) in environments of 6 mole % (top) and 12 mole % (bottom). The data are presented in the form of ffe, the mass of iron remaining in the coal/char, as a function of residence time. The mass of iron remaining in the coal/char has been normalized by its original amount, so that a value of 0.6 indicates that 40 % of the original iron in the coal has left the sample (see Equation 2). Both the rate and amount of iron release at particle residence times between 50 and 70 ms are high. Approximately half of the total iron in the char is released in this time interval. There is no release of iron prior to 45 ms and very little change in iron content after 70 ms. There are several indications of the uncertainty in the estimate of the iron loss. The scatter in the data points at anyone time represent the impact of the uncertainty in the measurement of m/mo on the extimated iron loss. This accounts for only a portion of the overall scatter in the data, however. The other portion is associated with variations in the amount of iron from sample to sample. This variation can be estimated either from replicated measurements or from other features of the experimental results . Replicated samples were prepared from the raw coal (residence time of o ms) and, in the 12 percent oxygen environment, at 95 ms and are illustrated in the figure. Also, the value of ffe reported at 47 ms in the 12 percent oxygen environment is about 1.17. Values of fi are physically limited to the range of 0-1. Therefore, this value of 1. 7 also gives some indication of the uncertainty in the iron measurement. Establishing confidence intervals for the data shown in Figure 2 is more difficult because the several methods of determining the overall loss are not entirely independent . That is , all of the methods depend on the same measurement of the percent iron in the sample. The remainder of the analysis for iron loss is independent, however. Some measure of the variation associated with estimating the amount of iron in each sample is needed. The mechanistic interpretation of these data, which is given later in this report , supports the experimental observation that no iron is lost during the coal devolatilization stage of combustion (the first 45 ms of residence time for the large particles). Therefore, the samples collected at times of 0 and 47 ms in each gas environment can be used to estimate the original content of iron and the variation associated with this measurement. Similarly, the variation in the last three data points of each section of the figure will be ascribed to statistical uncertainty. This should produce an estimate of the precision with which the iron loss is measured which is conservative. That is, the actual confidence intervals will be as small or smaller than 8 |
