OCR Text |
Show Sondreal and Ellman also applied regression analysis curve fitting of the previously discussed curves to bituminous and lignitic data. The correlation fits are summarized in Table 6. Schaefer's ratio offered the best fit for bituminous coals. None of the correlations intended for bituminous coals represented an applicable relationship for lignitic ash. Babcock and Wilcox22 has published the initial deformation and softening temperatures of bituminous coals as they relate to the percentage of iron in the ash for reducing and oxidizing conditions. The curves in Figure 5 dramatically show the impact of iron and environmental conditions on coal ash softening temperatures . Bryers and Taylor23 performed a step-wise, second-order, multiple regression analysis on eastern and western coals, using individual basic constituents as the independent variable. The program was designed to examine one variable at a time. It can select the most important variable first, develop a suitable expression, and then modify the expression to include the additional variables in a descending order of their significance. The resultant expression provides a very good fit to the data. However, the expression becomes too cumbersome to handle. Rather than fit the data with a good deal of precision, they decided to sacrifice accuracy and return to a single-variable, second-order regression analysis and use the percentage of basic constituents as the independent variable. Each rank coal was treated separately. For western fuels, separate correlations were developed for Si02/Al203 ratios approaching 1 and Si02/Al20a ratios much greater than 1. The results are summarized in Figure 6. As others 1-20 |