Unveiling the Hydrodynamics in Packed Beds of Non-Spherical Particles
Amirhossein Eghbalmanesh defended his PhD thesis at the Department of Chemical Engineering and Chemistry on April 11th.
Slender packed bed reactors are widely used in the chemical industry for processes involving catalytic conversions, such as the oxidative coupling of methane and the partial oxidation of ethylene. These reactors are often designed as multi-tubular due to significant heat liberation. However, these slender bed exhibit a low tube-to-particle diameter ratio, leading to pronounced non-uniform flow distribution. This, in turn, affects heat and mass transfer characteristics. In this thesis, a combined computational and experimental approach has been adopted to quantify this non-uniform flow distribution for both spherical and non-spherical particles.
A detailed experimental validation of PR-CFD results for packed beds with spherical and cylindrical particles is performed in this research using Magnetic Resonance Imaging (MRI) flow imaging. This experimental technique enables the non-invasive measurement of the velocity distribution in opaque systems such as packed beds of particles. Special attention has been paid to the removal of artefacts that are typically encountered in MRI flow imaging. These artefacts originate from the difference in magnetic susceptibility of the solid particles and the fluid and as a consequence, MRI faces challenges in accurately capturing the fluid velocity near the boundary of the solid particles. Nevertheless, it is shown that the agreement between results obtained from PR-CFD and improved MRI flow imaging is very good.
Furthermore, two numerical methods are discussed and employed to predict the flow distribution in packed beds with spherical particles. The first method is the validated Particle-Resolved Computational Fluid Dynamics (PR-CFD) method, which enables a prior prediction of the flow distribution and derived quantities such as the packed bed pressure drop. This method is extremely powerful but suffers from very high computational requirements. Therefore, a second method based on the Pore-Network-Model (PNM) is employed to predict the flow distribution and bed pressure drop at significantly lower computational costs. The results demonstrated that PNM produces accurate results for these quantities at a fraction of the computational cost of PR-CFD provided that careful calibration of the PNM hydraulic parameters is undertaken.
In addition, the simulation of the flow distribution in packed beds of non-spherical particles is reported. Such particles are frequently encountered in industrial practice and introduce additional complexities due to the presence of sharp edges and flat surfaces. Both PR-CFD and PNM have been used to predict the flow distribution and pressure drop in packed beds of cylindrical and cubical particles, where packed beds of spherical particles were taken as a reference system. It was found that the particle shape has a profound impact on the flow distribution and packed bed pressure drop. Regarding PNM, the pressure drop computed in PR-CFD was used to perform the necessary calibration of the PNM hydraulic parameters.
In conclusion, the effect of the packing configuration on the hydrodynamics of the packed bed reactors is investigated, and the results showed the significant effect of the packing density and particle shape on the pressure drop and fluid distribution.
Title of PhD thesis: Unveiling the Hydrodynamics in Packed Beds of Non-Spherical Particles - Particle-Resolved CFD modeling, Validation with MRI, and PNM Calibration Supervisors: Professor Hans Kuijpers and dr. Maike Baltussen and dr. Frank Peters.