Limitations Of Current Methodologies Used For Computational Fluid Dynamics (CFD) Model Development For Pulverized Coal Based Processes

The objective of this work is to investigate the inadequacies of current CFD based methodologies of model development for predicting combustion behavior of pulverized coal in furnaces. In the present paper, a steady state CFD model was used for in-depth analysis of pulverized coal combustion. The model accounts for turbulence, coal combustion, convective and radiative heat transfer and motion of coal particles. It also uses coal combustion models to account for important processes such as devolatilization, combustion of char and the evolution of fuel nitrogen during coal combustion. The model predicts fluid flow, heat transfer, progress of combustion and pollutant formation (NOx, etc.). The model uses several parameters, which need to be tuned with experimental data, which is not only limited in quantity but also insufficient in detail. Various existing methodologies like part calibration method for tuning of the parameters are discussed. In the present work, the parameters associated with coal combustion models were tuned and validated with published data obtained from an experimental furnace using one such part calibration methodology. The results clearly show that the model predictions differed with the experimental data not only quantitatively but also qualitatively. These discrepancies were attributed to part calibration methodology. A detailed parametric study was carried out to demonstrate the importance of parameters from various classes, namely, combustion kinetics, heat transfer, process and geometry parameters, and CFD model parameters. This investigation clearly indicates that parameters from different classes have significant effect on combustion behavior. The discrepancies in current methodology are discussed and a need for comprehensive methodology is highlighted. Keywords: pulverized coal combustion, computational fluid dynamics (CFD), pollutant formation (NOx), mathematical modeling