Liquid Holdup Measurements in Turbulent Bed Contactors z - - by a Tracer Technique zyx A. H. J. PATERSON and R. CLIFT* Department of Biotechnology, Massey University, Palmerston North, New Zealund An experimental technique to eliminate the effects of sensor dynamics when a tracer is used to measure the hydro- dynamics of a turbulent bed contactor zyxwvutsr is described. The resulting liquid holdup data are compared both with predictions from the pressure drop measurements and with the available correlations in the literature. The results show that the pressure drop method and Handl’s (1976) equation under-predict liquidholdup, while Rama et al.’s (1983) and Kito et al.’s (1978) equations over-predict the results. On dtcrit une technique exp6rimentale pour tliminer les effets de la dynamique du capteur lors de I’utilisation d’un traceur pour mesurer les parambtres hydrodynamiques d’un contacteur h lit turbulent. Les donnkes obtenues pour la rttention du liquide sont compartes aux prdictions faites zyxwv B partir des mesures de la perte de charge ainsi qu’aux corrtlations publites. Les rksultats montrent que la mtthode de la perte de charge et I’tquation de Handl (1976) sous- estiment la rttention du liquide, alors que les tquations de Rama et coll. (1983) et Kito et coll. (1978) la surestiment. Keywords: turbulent bed contactors, liquid holdup in TBC, tracer technique. turbulent bed contactor is a device used to contact a A gas with a liquid. It operates in the same manner as a packed column with the gas flowing upwards through the packing and the liquid flowing down. However, in a turbu- lent bed contactor, the packing is mobile, giving rise to its alternative name of a mobile bed contactor. The early work on turbulent zyxwvuts bed contactors introduced con- siderable confusion as different authors found that changing particle properties and gas and liquid flow rates influenced the minimum mobilisation velocity and total pressure drop across the bed in different ways (Adamiec et al., 1976; Askelrod and Yakovenko, 1969; Barile and Mayer, 1971; Blyakher et al., 1967; Davidson and Harrison, 1963; Chen and Douglas, 1968; Gel’perin et al., 1966b; Gel’perin et al., 1968; Handl, 1976; Tichy et al., 1972; Vetlugina et al., 1976). This confusion was dispersed by O’Neill et al. (1972) who identified four different flow regimes which can exist in countercurrent gas-liquid contactors. These regimes are shown schematically on Figure 1. For a constant liquid loading (&), the behaviour of the bed for increasing gas flow will be determined by the packing properties, the most important being the packing density. For gas flows insufficient to cause loading or fluidization of the packing, the bed behaves as a packed bed. This regime is marked AB on Figure 1. For a light density packing (subscript 1 on Figure 1) the minimum fluidization velocity (marked UGmr,) is reached before the flooding point and the bed fluidizes. This regime is called turbulent fluidization. Further increase in gas velocity does not increase the pressure drop across the bed (X1 to Yl). For a heavier packing (subscript 2) the flooding velocity indicated by the “flooding locus line” is reached before the fluidization velocity, (U,,,), and the packing floods with a corresponding increase in liquid holdup. The pressure drop increases accordingly until it counterbalances the weight of the liquid and packing. At this point, zyxwvutsrq X2, the bed will expand, allowing the liquid to flow past the packing and the bed will *Departmentof Chemical and Process Engineering, University of Surrey, Guildford GU2 SXM, U.K. then collapse. This process is repeated and has been named incipient flooding. Again, once the required pressure gradient is built up, it remains constant for further increase in gas velocities (X2 to Y2). The fourth regime described by O’Neill et al. (1972) was bubble flow through very dense, flooded, stationary pack- ing (subscript 3). The line at M on the diagram represents the line at which the column is completely flooded with the liquid forming the continuous phase. Lines L1, L2, and L3 represent increasing liquid flow rates. Turbulent bed contactors are operated in the turbulently fluidized or incipiently flooded regime. It is possible to change from incipient flooding to turbulently fluidized for a given packing density (subscript 4) by increasing the gas flow (X4 to Y4). When the previous literature is examined using the above classifications, the contradictory results of the effects of par- ticle properties and gas andor liquid flow rates on the mobili- sation velocity and pressure drop fall into different regimes. The effects reported do follow the expected trends for the regime of operation determined by the particle properties and fluid flow rates. Both of the above mechanisms of mobilisation predict that the pressure drop across a turbulent bed contactor counter- balances the combined weight of packing and liquid in the bed. Several authors (Gel’perin et al., 1966b; Groeneveld, 1967) have proposed that the liquid holdup in the column can be deducted from this relationship if the effects of the distributor plate are calculated separately. When this work was undertaken, three papers (Gel’perin et al., 1966b; Gel’perin et al., 1968; Groeneveld, 1967) had published work in which both the pressure drop and liquid holdup were measured simultaneously and independently. Gel’perin and co-workers measured the liquid holdup by simultaneously shutting off the gas and liquid flows and collecting the liquid left in the bed. The measured liquid hold- ups have been compared in Figure 2 with those predicted from the pressure drop data. This shows that the pressure drop method underpredicts the actual liquid holdup. Groeneveld (1967) used a step change in concentration of zy 10 THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING, VOLUME 65, FEBRUARY 1987