OCR Text |
Show decay is calculated using an average char particle size of 15 Jlm, which was adjusted to fit the data. Choice of this size results in a relatively rapid decay of the initial oxygen toward the local excess oxygen in each parcel of gas and char. The rate expression of Field et al. (1967), first order in oxygen, was used to describe the char-oxygen reaction, with its preexponential factor reduced by a factor of six, also adjusted to fit the measurements, as described below. Unburned carbon was found by integration of the expression for conversion of each size of char particles over the distribution of excess air, followed by integration of the carbon remaining at each size over the range of sizes of particles incompletely burned. The two parameters mentioned above were adjusted to make the calculations agree as closely as possible with the curve shown in Figures 2 and 7, derived from the correlation of carbon loss developed by Babcock & Wilcox (1963). The fit to the B&W curve was made for the condition of uniform excess air in the postflame region (<lEA = 0), on the assumption that the B&W correlation was based upon measurements in relatively new, tightly-controlled equipment, and on the expectation that the explanation for the higher carbon losses observed in boiler no. 12 might be associated with an increase in the width of the excess air distribution, due to inleakage of air. The calculated dependence of carbon loss on excess air is compared with the B&W correlation in Figure 7 (dashed curve at the far left). The calculation exhibits greater sensitivity to excess air than the B&W curve at low excess air, and predicts carbon loss increasing slightly with increasing excess air at high excess air, in contrast to B&W's experience. The latter effect is associated, in the calculation, with the decreases in residence time and furnace temperature with increasing excess air. These effects might be offset by the increase in the difference between char particle and gas temperatures as excess· air increases; an effect which was not included in the model. The calculated carbon loss increases markedly for values of excess air below zero (not shown in Figure 7), becoming roughly equal to the excess air deficit (for example, -10% excess air causes roughly 10% of the heat input to be lost as unburned carbon), until excess air is so low that no char is burned. A calculation of the dependence of carbon loss on average excess air for a distribution having a standard deviation of 3.5% excess air is shown in Figure 7 (dot-dash curve). In the range of excess air from 25 to 30%, the apparent condition of the tests, the calculated carbon 13 |