Choice of model for predicting the dispersion height in liquid/liquid gravity settlers from batch settling data

Abstract

The design of liquid/liquid gravity settlers has attracted increasing interest in recent years due to the industrial importance of solvent extraction. Batch settling tests are easier to perform than continuous pilot plant runs, but require sophisticated interpretation if the behaviour of a steady-state dispersion is to be predicted. Both batch and continuous dispersions contain sedimentation and dense-packed zones in which the drops grow in size through binary coalescence and finally disappear when they coalesce with their bulk homophase. The rate of droplet sedimentation is a function of the drop size and dispersed phase hold-up, whereas the interfacial coalescence rate is dependent on the height of the dense-packed zone and the drop size at the coalescing interface. At the steady state the dense-packed height may shrink to zero if the interfacial coalescence is virtually instantaneous. On the other hand, the whole of the dispersion may be dense-packed if the conditions in the feed stream lead to a volume rate of sedimentation which is higher than the dispersed phase throughput. Similar extreme behaviour also occurs in batch dispersions under corresponding conditions. The variations in height of the sedimentation and dense-packed zones, and hence the total dispersion height with the throughput in a continuous settler, can thus be predicted using parameters determined from experimental batch sedimentation and coalescence data. A procedure based on the shape of the batch decay curves, hold-up and volume rates of sedimentation and interfacial coalescence is proposed for selecting the appropriate model which is verified with available experimental data.