Simulations of large-scale coal combustors rely on accurate submodels to describe the chemical and physical changes in coal during reaction. Typically, simplified empirical submodels are tuned to experimental data to reduce the computational complexity. When data are not readily available, simplifying assumptions are used, which can create inaccuracies and biases in a large simulation. One such simplifying assumption in coal research is how to describe the elemental composition of primary pyrolysis products. This paper explores several different empirical model forms to predict the dry, ash-free fractions of C, H, O, N, and S in both the char and tar, using variables such as parent coal composition, reaction conditions (temperatures and particle residence times), and key coal structural parameters derived from NMR measurements to improve the treatment of coal chemistry in large simulations. These model forms were correlated to existing data using a wide range of experimental data using a cross-validation procedure. Since coal structural values can be expensive to measure, several correlations from the literature were used to estimate these values on the basis of information from the proximate and ultimate analyses of the parent coal, including a new correlation for the coal aromaticity. These model forms were tested against a set of measured elemental compositions of tar and char to find the best fit to use in the cross-validation process. The best empirical models are presented that predict the elemental composition of the coal char and tar after devolatilization.

The grate-kiln process is employed for sintering and oxidation of iron-ore pellets. In this process, a fuel (typically coal) is combusted with a large amount of excess air in a rotary kiln, and the high air-to-fuel ratio leads to significant NO x formation. The current Article is an assessment of NO x reduction measures that have been tested in pilot-scale and in full-scale by the Swedish iron-ore company Luossavaara-Kiirunavaara Aktiebolag (LKAB). The results show that the scaling between the full-scale kiln and the pilot-scale kiln is crucial, and several primary measures that reduce NO x significantly in pilot-scale achieve negligible reduction in full-scale. In the investigated full-scale kiln, thermal NO x formation is efficiently suppressed and low compared with the NO formation from the fuel-bound nitrogen (especially char-bound nitrogen). Suppressing the NO formation from the char-bound nitrogen is difficult due to the high amounts of excess air, and all measures tested to alter mixing patterns have shown limited effect. Switching to a fuel with a lower nitrogen content is efficient and probably necessary to achieve low NO x emissions without secondary measures. Simulations show that replacing the reference coal with a biomass that contains 0.1% nitrogen can reduce NO x emissions by 90%.

No abstract

No abstract

No abstract