Experimental and Numerical Analysis of Fluidization of Pelletized Activated Carbon Columns: Effects of Glass Bead Retention Layer

In pressure swing adsorption (PSA) processes there is always a desire to process as much gas as possible in the smallest beds possible. This necessarily leads to velocities that, especially during pressure-changing steps, may exceed the adsorbent particle fluidization velocity within the bed. To prevent the possibility of bed expansion and fluidization bed retention systems are usually employed at the top of the bed in the form of dense beads, bags of the same or springs, otherwise the mechanical integrity of the adsorbent might become compromised. PSA process simulation, based on Archimedes’ force balance involving weight, buoyancy and the drag force relative to the pressure drop predicted from Ergun’s equation, is a key method used to determine when the operating conditions may lead to fluidization conditions. However, recent bench scale experimental runs seem to agree with Archimedes’ balance only for a single layer packed bed, i.e., an unretained packed bed containing only the adsorbent. When placing a layer of glass beads on top of the adsorbent, the bed would fluidize only when the bed velocities were already several times those corresponding to that predicted by Archimedes’ balance. Most incredibly, the mass of glass beads used was only a fraction of the height of the absorbent bed. This significant discrepancy can only be assigned to friction exerted by the wall to keep the bed from moving, as described by the more than 100-year-old Janssen’s relationship To investigate bed expansion and fluidization under the influence of wall effects, a column with a full-scale height of 190 cm was constructed for operation under both ambient and elevated pressures. The bed consisted of a carbon-based material with the possibility of adding a layer of glass beads on top. It was exposed to progressively increasing flow rates until fluidization occurred, while maintaining a constant bed exit pressure. Parameters investigated included adsorbent height (0.24, 0.48, 0.73, 0.97 and 1.58 m), glass bead weight (0, 100 and 200 g), and bed pressure (20, 30, 40, 50 and 60 psia). The experimental results indicated that the minimum fluidization velocity increased with the weight of glass beads. Conversely, an increase in bed pressure and height led to a reduction in the minimum fluidization velocity. A modified form of Janssen’s equation, tailored to the experimental system, was developed to account for the frictional wall effects. This modification enabled the model to successfully capture the influence of the glass bead weight, bed height and outlet pressure.