Unresolved CFD-DEM modeling of multiphase flow in densely packed particle beds

摘要

Highlights•A VOF–DEM model was implemented in open-source softwares and used to study drainage in densely packed particle beds.•Instability issues inherent to the model were found for particle beds with particle densities lower than the fluid’s.•A smoothing approach was proposed to deal with these instabilities by filtering high frequency pressure fluctuations.•Experiments were conducted and the model successfully reproduced the measurements and observations.•Different drainage was observed when accounting for the particle bed dynamics of a demonstration blast furnace hearth.AbstractMultiphase flows in porous and granular media are widely encountered in many research and engineering fields. Due to the complexity and limited accessibility of real facilities, it can often be difficult to obtain information with measurements. Thus, accurate numerical models are desired tools for design and optimization purposes.In this work, a coupled CFD (computational fluid dynamics)–DEM (discrete element method) model is presented to study multiphase flows in densely packed beds. The well-known VOF (volume of fluid) method is used to describe an arbitrary number of continuous phases and DEM to model the disperse, solid phase. Since the porosity is directly calculated from the particle configuration, a dynamic, spatially resolved description of the granular medium is obtained in contrast to the widely used fixed-porosity-field Eulerian approaches.We discovered that artificial pressure fluctuations inherent to the CFD–DEM coupling method are critical in terms of stability when simulating particle beds with a solid-to-fluid density ratio less than unity. To make the calculations more robust, a novel smoothing model based on temporal relaxation was developed.Experimental validation was performed on an in-house measurement setup, where water was drained through sitting and floating beds, which we were able to successfully reproduce with our simulations.Finally, we demonstrate the capabilities and the advantages of our model by employing it on a complex multiphase application, the drainage of liquid iron and slag from a blast furnace hearth. A possible strategy to capture the large spatial and temporal scales of real plants is outlined for future work.