Preparative chromatography is attracting more and more interest from the fine chemicals and pharmaceutical industry, both for product development and commercial production. In the development of Life Science products, an important step is the choice and the design of cost-efficient unit operations for purification. Precursors for pharmaceuticals often have to be nearly 100% pure due to regulatory demands. Sometimes their physico-chemical properties differ little from those of the by-products, and they may be thermally unstable. In these cases, standard separations such as distillation are not applicable. Therefore chromatographic separation processes which can be operated at low temperatures are increasingly used for preparative separations of products in the food, fine chemicals and pharmaceutical industries. The classical chromatographic separation process is batch chromatography where a column is charged with pulses of the feed solution and the components separate as the mixture travels along the column. A more effective variant with respect to solvent consumption and utilisation of the adsorbent is SMB (simulated moving bed) chromatography where a number of columns (usually 6-16) are arranged in a closed loop and a continuous separation is achieved by switching the input and output ports. The price for the higher effectiveness of the SMB process and its variants is their considerably higher complexity and sensitivity to the operating parameters which makes the choice of the plant setup and the operating conditions a challenging task and necessitates close supervision and control if the process is operated near the economic optimum. Due to the distributed, nonlinear character of the processes, chromatographic separations call for the use of advanced control concepts and on-line optimization We report on the results of a long-term research effort on the operation of chromatographic separations which has been performed in close co-operation with the plant design group in our department. First a rigorous model-based approach to the optimization of the configuration and the operating parameters of continuous chromatographic separations is presented, including the newly introduced VARICOL process where the ports are switched asynchronously. Then a “classical” control approach is discussed where the optimal operation point characterized by the positions of the concentration fronts is stabilized against disturbances by feedback control. This however cannot assure optimality if process parameters change. Therefore a direct optimization-based control concept was developed where the desired product purities do not appear as control objectives but as constraints while an economic cost function is minimized on-line. The effectiveness of this concept is demonstrated by experiments at a pilot plant for reactive chromatography. A key problem in model-based optimization and operation of chromatographic processes is the need for an accurate process model. It is usually very time-consuming and costly to determine the crucial parameters of the process, in particular the adsorption isotherms, experimentally. In our MPC concept, this problem is solved by on-line re-estimation of the key parameters of the process model. We demonstrate that in SMB processes, the information which is gained during operation is sufficient for good estimations. In batch chromatography, the situation is less favourable so that there is a strong interest in on-line optimization techniques which work without an accurate process model. A suitable approach is proposed.

Sebastian Engell received his Dipl.-Ing. Degree in Electrical Engineering from the Ruhr-Universitaet Bochum, 1978. He also holds a Dr.-Ing. degree in Control Theory from the Universitaet Duisburg, 1981. He is currently a Professor of Process Control at the University of Dortmund, Department of Chemical Engineering. His research interests concern hybrid systems, especially modelling, simulation and optimization, and control systems design. He is the Co-Editor of the IEEE Transactions on Control Systems Technology and Associate Editor of the European Journal of Control.