Computational investigation of oxy-combustion of pulverized coal and biomass in a swirl burner

Swirling pulverized coal and biomass flames are computationally investigated for oxy-combustion. The two-phase flow is described by a Eulerian-Eulerian approach. For radiation, the absorption coefficient is approximated by superposing particle and gas contributions, considering oxy-combustion conditions for the latter. Turbulence is modelled within a URANS framework, using the standard k-epsilon model and Rey-nolds Stress Model (RSM). It is observed that RSM captures the unsteady dynamics of the coherent structures, whereas they are not captured by k-epsilon model. Predicted velocities are compared with mea-surements. It is observed that the RSM predictions are in a better agreement with the measurements compared to the k-epsilon model. The discrepancy between the predictions and measurements can most clearly be quantified in terms of the peak values of the axial velocity in the forward flow region enveloping the inner recirculation zone. The calculations constantly underpredict the measurements. On the average, this is about 32% for the RSM and 52% for the k-epsilon model, for both flames. The biomass flame is predicted nearly twice as long compared to the coal flame. As means of verification, the coal flame is additionally calculated using a classical Eulerian-Lagrangian two-phase formulation, leading to quite similar results to the Eulerian-Eulerian formulation. (c) 2021 Elsevier Ltd. All rights reserved.

Automated summary

Computational investigation of swirling pulverized coal and biomass flames for oxy-combustion reveals that the Reynolds Stress Model (RSM) captures unsteady dynamics better than the k-epsilon model. Predicted velocities show better agreement with measurements using RSM. However, both models underestimate the measurements, with discrepancies quantified in peak values of axial velocity. Biomass flame is predicted to be nearly twice as long as coal flame.