(62c) Effects of Grid Topologies and Resolution of Boundary Layer on Gas-Particle Flow Predictions in Circulating Fluidized Bed Riser | AIChE

(62c) Effects of Grid Topologies and Resolution of Boundary Layer on Gas-Particle Flow Predictions in Circulating Fluidized Bed Riser

Authors 

Li, T., National Energy Technology Laboratory
Shahnam, M., National Energy Technology Laboratory
Sundaresan, S., Princeton University

As a part of larger study to test the filtered Two-Fluid models (TFMs) [1,2] for gas-particle flows in circulating fluidized bed (CFB), we established the baseline results of kinetic theory based TFM by performing simulations of CFB riser using different grid resolutions and grid topologies. The specific cases simulated are based on CFB challenge problems involving Geldart Group A particles put forward by the National Energy Technology Laboratory (Morgantown, WV) and the Particulate Solids Research Inc. (Chicago, IL). All simulations were carried out using ANSYS Fluent 14.  Time-averaged gas-phase axial pressure gradient and particle-phase vertical velocity results from the simulations were compared with the experimental data. Cut-cell and hybrid grids that did not include boundary layer cells manifested apparent convergence upon successive grid refinements, but failed to predict downflow of solids near wall and yielded significantly lower time-averaged solids holdup than the experiments. Upon addition of boundary layer cells near the riser wall, both grids predicted downflow of solids but the results remained dependent on boundary layer resolution. In fact, with sufficient boundary layer resolution, the predicted solids holdup could be closer or even higher than the experimental data but it always coincided with overprediction of particle-phase vertical velocity near riser wall as well as in the core region.

In summary, we found that

-          Presence of boundary layer grid is essential for predicting downflow of solids near riser wall.

-          Different grid topologies yielded nearly similar results (except some localized differences) when boundary layer was sufficiently resolved.

-          Results remained dependent on grid as well as boundary layer resolution even after successive refinements, which confirmed the need for filtered TFMs to improve gas-particle flow predictions in riser.

  1. Y. Igci and S. Sundaresan, Ind. Eng. Chem. Res., 50, 13190–13201 (2011).
  2. C.C. Milioli, F.E. Milioli, W. Holloway, K. Agrawal, S. Sundaresan, AICHE J., 59,3265-3275 (2013).