Analyzing the effects of turbulence and multiphase treatments on oxy-coal combustion process predictions using LES and RANS

Numerical studies of pulverized coal swirl combustion in oxy-fuel atmosphere are carried out. Thereby two issues are especially addressed: (1) how LES and RANS impact differently the predictions of combustion properties even though, in both approaches, the same kinetic rates are used to represent the coal combustion processes; (2) how the numerical multiphase treatments may affect the prediction of micro-process interaction as well as the range in which these processes are not negligible. For that purpose a methodology is developed based on an Eulerian-Lagrangian oxy-coal combustion module which is designed relying on the state of the art models as implemented in the commercial code ANSYS Fluent 17. This especially includes three kinetic rates for the description of coal combustion, namely coal devolatilization, volatile combustion and char combustion. Based on an appropriate Stokes number consideration, a full two-way inter-phase coupling has been numerically adopted. To assess the prediction capability of the overall model, a new set of experimental data from a 60 kW(th) oxy-coal test facility is employed. First, the model validation is ensured by comparison of results in terms of flow field and products from volatile and char combustion. Then, an analysis is performed to elucidate how the two-phase turbulence modeling impacts the thermal flow predictions along with the evolution of multiphase oxy-coal combustion properties. Finally, it is demonstrated how the numerical multiphase treatments affect the prediction of micro process interaction in terms of coal devolatilization, coal particle distribution due to turbulent particle dispersion, and of gaseous heat release as well as char burnout. The range in which these interphase processes (subgrid scale particle dispersion) are not negligible is also pointed out in terms of subgrid scale Stokes number. (C) 2017 Elsevier Ltd. All rights reserved.