Comparison between pseudohomogeneous and resolved-particle models for liquid hydrodynamics in packed-bed reactors

Abstract

Despite the advances in the modelling of Packed-Bed Reactors (PBR) it is common to find in literature works where resolved-particle models are implemented considering small-scale domains; or works where large-scale models are implemented with a pseudohomogeneous approach. In order to quantify the differences in the prediction of the local velocity between these two approaches, a CFD study was conducted for the modelling of the single-phase flow in a PBR. Six different cases, two resolved-particle and four pseudohomogeneous, were implemented and compared. The resolved-particle cases incorporated the explicit description of the solids bed distribution, considering an ordered and a randomly distributed bed of spheres, in which the punctual momentum balance was solved in the interstitial void space left by the solids. The pseudohomogeneous cases considered Darcy’s law (DL) and the Navier-Stokes-Darcy-Forchheimer (NSDF) equation, and the bed porosity as an overall average value εB=εB and as a function of the radial position of the bed εB=εBr, in order to include the effects of the solids bed without the explicit description of the solids distribution. The results show that there are significant local differences in the textural characteristics between the random packing and ordered packing for the resolved-particle cases. It was observed that the pseudohomogeneous models allow to predict the macroscopic behaviours indicated in the resolved-particle case, but fail to capture the local behaviours inside the bed. Furthermore, the inclusion of an average porosity distribution function does not improve the prediction of the local scale phenomena.