Simulating coarse particle conveying by a set of Eulerian, Lagrangian and hybrid particle models


B Popoff, C Kloss, D Kahrimanovic, M Braun, S Pirker




Powder Technology



This paper considers the behaviour of mono-disperse coarse glass particles in a pneumatic conveying experiment which consists of a curved and straight rectangular duct. Thus, the flow situation comprises the formation of a particle strand in the curved section as well as its dispersion in the subsequent straight horizontal channel. Thereby, from a physical point of view the effects of inter-particle collisions, particle rotation and wall roughness are of crucial importance.Based on this experimental setup six numerical models for the granular phase are applied in order to picture these physical phenomena. While the first set of three models - the (a) Lagrangian Discrete Phase (DP) model, the (b) Discrete Element Method (DEM) and the (c) 'standard' Eulerian-granular (EUgran) model - is readily available in commercial codes the remaining three models represent in-house developments. As a first modification the standard DP model is enhanced by sub-models accounting for inter-particle collisions, wall roughness and particle rotation in order to get an (d) enhanced Discrete Phase (DP+) model. Next, two combinations between the Eulerian and Lagrangian models, the (e) Dense Discrete Phase Model (DDPM), and a (f) Eulerian based hybrid model (EUgran+) are presented and discussed.Thus, all in all six numerical models are evaluated by qualitatively checking the main flow pattern, subsequently by a quantitative validation of dedicated profiles of particle velocity and concentration and finally, by qualitatively comparing the computational effort of each numerical model. While all 'out of the box' models fail in predicting even basic flow patterns the remaining enhanced models agree well with the experimental results. Nevertheless, their required computational effort is significantly different.


Coarse particle conveying, Discrete Element model, Eulerian model, Hybrid model, Lagrangian model

Access Full Text