No abstract

Liquid-phase Peclet numbers have been measured for !h in. Raxhig rings with countercurrent air flow in a 2 in. 1.0. column. The response curve has been analyzed by a computer program, since the output doto was collected directly on paper tape. The Peclet number calculated from a moments analysis is considered unreliable, and the diffusion model can be rejected for the liquid phase. If the Peclet number is calculated by a least-squares fit, and if the result is forced through the peak in the response curve, then a correlation for over one hundred runs with got flows from 0 to 99% of flooding is given by Equotions (7) and (8).The packed column with two-phase flow is widely used as an item of mass transfer equipment, but the flow pattern in both phases is poorly understood. The liquid phase often flows by gravity through the bed and is described as trickle flow. The liquid path is probably a large number of laminar films which, after a passage through the bed, are partly transferred to the wall. When the ratio of column diameter to nominal packing size exceeds 8 to 1, there is only a small fraction of liquid on the wall after several column diameters, even for a single point liquid distributor. A study of the residence time distribution in the liquid phase shows a considerable departure from plug flow, and the dispersion is often characterized by a dispersion coefficient. Studies of this dispersion should lead to an improved understanding of the mass transfer operation.There has been reported in the literature a large number of experiments on dispersion in packed beds with singlephase flow. The published results on two-phase %ow include the gas-phase dispersion measurements of DeMaiia and White (15) A large number of models have been examined to characterize the dispersion by comparing the input, output responses of the bed. With single-phase flow, a diffusion model has been satisfactory in modeling the output response for pulse, step, and sinusoidal input signals, although Carberry and Bretton (1) have noted some anomalous behavior. The situation is not so clear with two-phase flow. Hoogendoorn and Lips (7) and Otake and Kunugita (9) have found that the diffusion model is unsatisfactory owing to the output response hat;ing a long tail for a pulse input signal. Dunn et al. ( 3 ) have used a step change for an input, and they found the diffusion model did not fit the output response exactly, although it was the best model examined. Sater and Levenspiel (11) reported long tails on two output responses at different positions along the bed for a near pulse input signal.The output signals mentioned have been recorded on a strip chart, and the lines fluctuate about a mean position both for a radioactive sensor used by Sater and Levenspiel E X P E R I M E N T A L APPARATUSA 2 in. I D . glass column was packed to a depth of 5 ft.with y4 in. ceramic Raschig rings, as shown in Figure 1. Water was pumped from a reservoir through a rotameter to a liquid distributor. Air was passed through a rotameter and entered below the pac...