Physically meaningful modeling of the 3-zone and 4-zone simulated moving bed processes

This al-tide is intended as a complement to a recent article in this journal, ''Comparative Study of Flow Schemes for a Simulated Countercurrent Adsorption Separation Process'' by Ching et al. (1992). These authors carried our experiments on continuous separation of three carbohydrate mixtures with both the usual four-section scheme and a three-section scheme. fructose dextran (MW congruent to 9,400), raffinose dextran (MW = 6,000), and fructose-raffinose, These two schemes were modeled as equivalent true countercurrent systems. From the experiments and model results they conclude that the equivalent true countercurrent model is fully adapted to represent the four-section scheme, but a major discrepancy arises with the three-section scheme: the raffinate concentration is not correctly predicted. From these results it may be concluded that when a physically meaningful modeling of the equivalent true countercurrent is realized, the apparent discrepancy disappears. From another standpoint, such a mismatch between the experiments and model predictions would nor have been observed with a model that considers the periodic port motion. Such a model foresees a cyclic steady state where extract and raffinate compositions vary throughout the period T; they can be considered as constant only when averaged over the period. It is worth comparing the results of these two models, especially when the equivalent countercurrent representation is not very close to the physical phenomenon it should represent. We agree with their major finding that a three-zone scheme represents necessarily a higher desorbent demand than a four-zone scheme.