Unburned carbon loss from pulverized coal combustors

Abstract

Abstract Because the amount of unburned carbon in fly ash from pulverized coal-fired boilers is such a small proportion of the combustibles in the fuel, its origin is usually in atypical properties of the fuel and furnace, rather than their average characteristics. Knowledge of the distributions of coal properties and combustion conditions is, therefore, one of the requirements for a useful simulation of carbon burnout. A reactor engineering approach to this problem, which accounts for the distributions of particle size, residence time, and mixedness of fuel and air has been adopted. The model was tested by comparison with measurements of combustibles in fly ash from a 136000 kg steam h−1 boiler firing medium (mvb) and high (hvb) volatile bituminous coals. Size distribution, as well as the concentration of unburned carbon in fly ash were used to test the accuracy of the simulation of char combustion. The key to the calculation is the allowance for distributions of residence time in the furnace and mixedness of coal and air, without which it is not possible to explain the presence of small unburned particles in fly ash. The coal and plant data required for the calculations are: proximate and ultimate analyses, heating value, screen analysis of the pulverized coal, the dimensions of the furnace, excess oxygen, air preheat, furnace exit gas temperature and a representative sample of the fly ash. The fly ash samples were screened and the carbon content determined for each size cut. The model reproduced the observed dependence of amount and size distribution of unburned combustibles on both grind and volatile matter. The principal adjustable parameters in the calculation are the rate coefficient for char combustion, the mixedness of fuel and air at the end of the volatiles flames, the characteristic mixing time in the post-flame region, and the average gas temperature in the furnace. Although no particle-to-particle variation of char reactivity was introduced, the slow char combustion rate required to fit the measurements suggests that unreactive char may have made a disproportionate contribution to the unburned. The model agrees with the dependences of unburned combustibles on volatile matter, furnace loading and coal fineness reported by Babcock and Wilcox. However, the model exhibits higher sensitivity to excess air than the Babcock and Wilcox correlation. The accuracy of the calculation might be improved by introducing a lower than first-order dependence of the char combustion rate on oxygen and by accounting for size reduction of char particles during burnout.