Design and optimisation of a simulated moving bed unit: role of deviations from equilibrium theory

The design of a simulated moving bed involves thermodynamic, kinetic and hydrodynamic aspects and requires the optimisation of several variables: plant design variables, such as the column length and diameter, and operating variables, among them four independent flow-rates, the feed concentration and the switch time. In this work we develop an algorithm to design both the unit and its operating conditions, with an overall view on equilibrium properties, efficiency and hydrodynamics, using a simple equilibrium stage model. In this way we determine the parameters leading to the highest possible productivity for a given separation, only requiring the knowledge of the equilibrium isotherms, the Van Deemter equation and a correlation for pressure drop. The algorithm has been used to investigate the effect on the separation performance of some parameters, such as particle size and required product purity, which are not considered by equilibrium theory. The results have been compared with the predictions of equilibrium theory and the observed deviations have been put in evidence and discussed.