OCR Text |
Show The imposition of kinetic constraints on our calculations can change the chemistry in important ways. We present one example of an expected kinetic constraint. Our computer calculations, when coupled with an experimental program, will permit us to deduce kinetic constraints and will, ultimately, lead to reliable predictions of combustion chemistry. Thus, our method should greatly reduce the number of measurements needed to deduce the chemistry for any particular coal and will aid in formulating strategies for reducing fouling and corrosIon. CALCULATIONS Calculations were performed for the combustion of lllinois #6 coal under oxidizing conditions (SRO --1.1) with a composition given in Table I. The major difference between these calculations and our prior work is in the use of a different method for representing the solution chemistry of molten silicates. The silicate model used here provides a very good representation of the thermodynamic properties of molten silicates.3- 8 The method is based on a modification of the quasichemical theory and permits us to simultaneously fit all thermodynamic data on a given binary silicate system to a set of modified quasichemical equations that can simulate the properties of ordered liquids. The resultant energy parameters (seven or less, including temperature coefficients) represented all available data (six binary silicate systems r very well. In addition, we have developed a combining rule that permits us to predict the solution properties of multicomponent silicates from those of the subsidiary binaries. It has been demonstrated that these predictions are in good agreement with measurements on ternary systems and are consistent with the best available theories (which apply for basic solutions). This lends confidence in the equations ability to predict the solution properties for multicomponent silicates. Another difference from our prior work is that the modified quasi chemical equations have been coupled with a new robust and powerful version of SOLGASMIX that permits one to perform calculations on large complex systems. In this system, we considered a slag phase containing Si02, Na20, FeO, CaO, and MgO, a sulfate phase containing Na2S0., CaSO., and MgSO., a gas phase and all possible solid phases. In reality, the sulfates should dissolve to some extent in the silicate and the silicate in the sulfate. We have not considered such solutions, which would require developments beyond the current state of the art. The result of such considerations would be to extend the range of stability of the silicates to lower temperature and of sulfates to higher temperature. We have developed theories capable of calculating the properties of such solutions and hope to incorporate them in our calculations in the future. 9- 12 In addition, we have considered the liquid sulfates to be an ideal solution and assumed that there were no sulfate solid solutions. This neglect probably makes the liquid stable to somewhat lower temperatures than if we had performed a complete analysis of data for sulfates and taken the known solid solutions into account. Thus, we have cut off our calculations at 1050 K since solid solutions of e.g., CaS04 with Na2S04 should form at lower temperatures. We plan an independent analysis of available data on sulfates that will be incorporated into the calculation. In addition, we have not considered the plagioclase solid solutions and have only considered the end member silicates, anorthite (CaAI2Si208) and albite (NaAISi308). We also plan to incorporate such solid solutions into our calculations. However, because plagioclase is a relatively acid silicate with a complex structure, it is hard to form the solid solution and one expects a kinetic con~traint on its formation. In addition, 2 |