Validation of DigiPac Simulations of Packed Columns Using X-ray Microtomography

X.Jia1, C. Xu1, R. Caulkin1, R. A. Williams1, H. Stitt2, M. Nijemeisland2, S. Aferka3, M. Crine3, A. Léonard3, D. Toye3 and P. Marchot3

1. Institute of Particle Science and Engineering, School of Process, Environmental and Materials Engineering, University of Leeds, Leeds, LS2 9JT, UK, Email: X.Jia@leeds.ac.uk

2. Johnson Matthey Catalysts, PO Box No 1, Billingham, Stockton on Tees, TS23 1LB, UK

3. Chemical Engineering Laboratory, Institut de Chimie, B6c, Université de Liège, Sart Tilman, B4000, Liège, Belgium

ABSTRACT

Modern X-ray microtomography is a non-destructive 3D imaging technique. It provides higher resolution structural information faster and cheaper than traditional stereological techniques, and thus is an excellent measurement technique to use for validation of simulation models of particulate structures. A novel particle packing algorithm, named DigiPac, was proposed earlier. In its original form, particles are represented digitally as coherent collections of 3D pixels, and their packing is simulated using a probabilistic approach with no explicit consideration of the physical forces involved. While it has been demonstrated to give adequate predictions of the poured bulk packing density in several real-world case studies involving complex shaped particles, its application to packed columns proved to be less satisfactory. It was clear that, for packed columns, the effects of inertia and collision forces could not be neglected.

Two new versions of the DigiPac algorithm have thus been developed to incorporate particle interactions explicitly. To distinguish them from the original version, they are named DigiDEM and DigiCGP. DigiDEM is an implementation of the Discrete Element Method for digitally represented particles, which treats particle interaction forces in a rigorous manner. DigiCGP stands for digital collision-guided packing algorithm. It treats particle interactions less rigorously, for the sake of speed. The two new versions have been used to simulate packed beds of different pellet shapes in containers of different geometry. To validate the simulation models, X-ray CT scans of the packed beds have been obtained.

This paper presents and compares some of the simulation and CT results, in terms of bulk density, density distribution, and pellet orientations. It is concluded that for packed columns the incorporation of physical interactions into the packing algorithm, even at a superficial level, can bring about a marked improvement in the predicted particle-level packing structures, compared with the original version. For packed columns, DigiCGP is useful for quick and qualitative trend finding, while DigiDEM can provide quantitatively more reliable information.

Keywords X-ray tomography, packed column, particle shape