A CFD–DEM study of the behaviour of single-solid and binary mixtures in a pyramidal spouted bed

作     者：Filippo Marchelli Cristina Moliner Barbara Bosio Elisabetta Arato 

作者机构：Faculty of Science and TechnologyFree University of Bozen-BolzanoPiazza Universita 539100 BolzanoItaly Department of CivilChemical and Environmental EngineeringUniversity of GenovaVia all'Opera Pia 1516145 GenovaItaly 

出 版 物：《Particuology》 (颗粒学报（英文版）)

年 卷 期：2019年第17卷第1期

页      面：79-91页

核心收录：

学科分类：08[工学] 0801[工学-力学（可授工学、理学学位）] 

基　　金：The authors are grateful to Professor Giorgio Rovero and Doctor Massimo Curti for the development of the studied spouted bed  and the contribution during the experimental campaign for the validation of the model 

主　　题：Spouted beds Computational fluid dynamics Discrete element method Binary mixtures Segregation phenomena 

摘      要：We simulated a lab-scale cold-flow spouted bed through computational fluid dynamics (CFD), coupled with the discrete element method (DEM) for the solid phase, using a commercial CFD program, ANSYS FLUENT 18.0. To limit the computational expense, we tested both a simplified pseudo-2D geometry and a complete 3D geometry. We found that the Haider and Levenspiel drag model is suitable for the pseudo-2D geometry;however, this model does not correctly predict fluidisation in the 3D geometry. Conversely, the Gidaspow drag model behaves accurately in the 3D geometry but overestimates the motion of particles in the pseudo-2D geometry. We studied several single-solid and binary mixtures to assess the reproducibility of segregation phenomena. The pseudo-2D model was able to predict the onset and minimum spouting flow rates of all mixtures with good accuracy. An analysis of the volume fraction contours of the binary mixtures permitted us to confirm that segregation phenomena were correctly predicted at low gas velocities. We showed that segregation decreased as the inlet gas flow rate was increased. Calculations performed in the complete 3D geometry were preliminarily assessed as more reliable but required almost four times as much computational time as those for the pseudo 2D geometry.