Premium
NMR velocimetry of flow in model fixed‐bed reactors of low aspect ratio
Author(s) -
Ren Xiaohong,
Stapf Siegfried,
Blümich Bernhard
Publication year - 2005
Publication title -
aiche journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.958
H-Index - 167
eISSN - 1547-5905
pISSN - 0001-1541
DOI - 10.1002/aic.10318
Subject(s) - dimensionless quantity , flow (mathematics) , spheres , aspect ratio (aeronautics) , tube (container) , mechanics , porosity , dispersion (optics) , flow velocity , velocimetry , chemistry , volumetric flow rate , materials science , geometry , nuclear magnetic resonance , physics , optics , composite material , mathematics , astronomy
The velocity distributions of flow of a single fluid phase through packed beds of various geometrical properties were investigated by combining magnetic resonance imaging (MRI) with velocity‐encoding and pulsed‐field‐gradient nuclear magnetic resonance (PFG‐NMR) experiments. The beds were generated from random packings of spherical glass beads and commercial porous catalyst pellets with spherical and cylindrical shape of different sizes d p inside cylindrical tubes of diameter d t . Flow investigations were performed for reduced dimensionless tube diameters d t /d p (aspect ratio) in the range 1.4–32. The influence of pellet ordering effects on the distribution of flow channels is demonstrated using static and velocity encoded spin density images. The velocity distribution averaged over the radial coordinate in the tube follows an oscillatory pattern that largely reflects the ordering of the particles themselves. This is found for spheres and, although less pronounced, for irregularly shaped catalyst particles. In all cases, flow occurs mainly along a few backbone branches, constituting a small fraction of the total cross section of the reactor. With the exception of large aspect ratios, the dominating contribution to flow occurs near the inner tube wall. The evolution of the relative fractions of flowing and stagnant fluid is discussed as a function of encoding time, and time‐dependent dispersion coefficients are obtained for the preasymptotic case. © 2005 American Institute of Chemical Engineers AIChE J, 51: 392–405, 2005