| description abstract | This paper presents computational simulations of flows in packed beds and compares the computational pressure-drop results with those given by the Ergun correlation. The computational methodology used in this work follows the combined discrete element method (DEM) and computational fluid dynamics (CFD) technique. DEM is used to predict the locations and packing structure of the particles in the bed, while CFD is used to predict the flow field in the void space surrounding the packed particles. The computational results obtained for irregular packed beds show that the local packing-structure parameters have significant effects not only on the local velocity and pressure fields but also on macroscopic quantities, such as the average pressure gradient along the length of the packed column. The computational results also show that classical correlations based on averaged values, such as the Ergun correlation, have poor predictive accuracy for macroscopic variations along a packed column, and this is mainly because such correlations do not account for local packing-structure parameters. The computational results confirm the existence of sections with linear variation of macroscopic parameters along the length of the packed column, and this leads to the conclusion that accurate results from DEM-CFD methods on shortened columns can be extrapolated to full-length columns. Moreover, it was found that unlike regularly packed beds, the predicted pressure for randomly packed beds experiences an apparent strong recovery near the downstream end of the packed bed, and then experiences a strong dip down to the plateau leading to the exit pressure. | |
