Effect of Tube-to-Pellet Diameter Ratio on Turbulent Hydrodynamics in Packed Beds: A Magnetic Resonance Velocity Imaging Study

Acknowledgements: The authors thank Shell Global Solutions International B.V. for funding this work, provision of the catalyst pellets used in this study, and providing water absorption measurements for determining pellet material density. SVE further thanks the Sir Winston Churchill Society of Edmonton and the Natural Sciences and Engineering Research Council of Canada (NSERC) for additional funding. For the purpose of open access, the authors have applied a Creative Commons Attribution (CC BY) licence to any Author Accepted Manuscript version arising from this submission.Funder: Sir Winston Churchill Society of EdmontonFunder: Shell Global Solutions International; doi: http://dx.doi.org/10.13039/100016773AbstractThe hydrodynamics in packed reactors strongly influences reactor performance. However, limited experimental techniques are capable of non-invasively measuring the velocity field in optically opaque packed beds at the turbulent flow conditions of commercial relevance. Here, compressed sensing magnetic resonance velocity imaging has been applied to investigate the hydrodynamics of turbulent flow through narrow packed beds of hollow cylindrical catalyst support pellets as a function of the tube-to-pellet diameter ratio, $$N$$ N , for $$N=$$ N = 2.3, 3.7, and 4.8. 3D images of time-averaged velocity for the gas flow through the beds were acquired at constant Reynolds number, $$R{e}_{\mathrm{p}}=$$ R e p = 2500, at a spatial resolution of 0.70 mm ($$\tt x$$ x ) $$\times$$ × 0.70 mm ($$\tt y$$ y ) $$\times$$ × 1.0 mm ($$\tt z$$ z ). The resulting flow images give insight into the bed and pellet scale hydrodynamics, which were systematically compared as a function of $$N$$ N . Some changes in hydrodynamics with $$N$$ N were observed. Namely, the near-wall hydrodynamics changed with $$N$$ N , with the $$N=$$ N = 4.8 bed showing higher velocity at the wall compared to the $$N=$$ N = 2.3 and $$N=$$ N = 3.7 beds. Further, in the $$N=$$ N = 3.7 bed, channels of high velocity, termed flow lanes, were found 1.3 particle diameters from the wall, possibly due to the bed structure in this particular bed. At the pellet scale, the hydrodynamics were found to be independent of $$N$$ N . The results reported here demonstrate the capability of magnetic resonance velocity imaging for studying turbulent flows in packed beds, and they provide fundamental insight into the effect of $$N$$ N on the hydrodynamics.