State Estimation in Chromatographic Smb Processes with Linear Adsorption Isotherms

Chromatographic separations are widespread in the fine chemicals and pharmaceutical industries for the downstream processing of complex mixtures with similar thermodynamic properties. In recent years, continuous chromatographic processes, in particular Simulated Moving Bed (SMB) processes are increasingly applied in industry. SMB processes are characterized by a counter-current movement of the liquid and the solid phase which is achieved by switching the inlet and the outlet ports of a closed loop of chromatographic columns periodically. Recent research has focused on the model-based optimization and control of SMB processes. In a model-based approach, plant/model mismatch has to be taken into account. In this paper, the issue of state estimation in SMB processes under plant/model mismatch is addressed. A Kalman filter is designed and successfully applied. The estimation is based on only one measurement device located in the recycle stream of the process. Copyright c ©2005 IFAC