Optimization of Ternary Simulated Moving Bed Separation by Superstructure Formulation

Simulated moving bed (SMB) chromatography has been widely studied in past and there have been various operating schemes proposed for efficient separation of binary mixtures. SMB is a promising option of separation because of its capability to achieve efficient separation even when the components differ very little in terms of affinity towards the stationary phase. However, separation of a ternary mixture using SMB is still a major drawback of this technology. Although many different strategies have been proposed, previous studies have rarely performed comprehensive investigations for finding the best ternary separation strategy from various possible alternatives. Our study addresses this issue by adopting a systematic nonlinear programming approach. A full superstructure formulation has been proposed which encompasses numerous operating schemes proposed in the literature for ternary separation. The optimization problem, constrained by partial differential algebraic equations model, is fully discretized both in temporal and spacial domain. The resulting nonlinear programming problem is handled by IPOPT 3.0, an interior point solver. These results are obtained for linear adsorption isotherms considered in Mata and Rodrigues (2001). We demonstrate that the superstructure approach has a potential to find more advantageous operating scheme than the existing operating schemes.