Stylianos Kyrimis, Robert Raja, Lindsay-Marie Armstrong

Abstract: Computational Fluid Dynamics (CFD) models are a valuable tool for the design, optimization, and scaling-up of fixed bed chemical reactors. However, the realistic representation of the catalytic bed structure and the mesh quality of the 3D geometry is of paramount importance to improve the accuracy of CFD models. For the former, computed tomography (CT) is a non-destructive method to map and generate the internal structure of experimental fixed bed reactors, enabling a direct 1-to-1 coupling between experiments and simulations. In our previous work, the internal structure of highly poly-dispersed fixed bed reactors, formed by sieved particles, was analysed. The particles that formed them displayed a wide range of sizes, shapes, and orientations. Due to the local topological complexity of these beds, meshing and simulating their entire volume would lead to exhaustive computational demands. To reduce these, a suitable sample section should be selected, which accurately represents both the bulk and the radial porosity of the full bed. Three distinct sample sections were quantified here for their accuracy, identifying that, due to the highly heterogeneous nature of the full beds, sample selection is case sensitive. In addition, compared to smaller particles, larger particles form more heterogeneous local structures, thus requiring longer sections to accurately represent the full bed. A selected 10% section was then meshed, and its hydrodynamic profile resolved, to evaluate its mesh independency. The results highlight the importance of choosing a suitable bed section and mesh size to reduce the computational demands, minimise the computational errors, and achieve the desired level of solution detail.

Keywords: Computational Fluid Dynamics (CFD), polydispersed beds, fixed bed chemical reactors, computed tomography (CT), catalytic particles, SAPO-34

Date Published: August 8, 2023
DOI: 10.11159/jffhmt.2023.004

View Article