Note: This content is accessible to all versions of every browser. However, this browser does not seem to support current Web standards, preventing the display of our site's design details.


Online optimization and control of simulated moving bed processes


C. Grossmann

Diploma/Master Thesis, SS 05

The increasing number of chiral drugs and biotechnological products in everyday life are subject to strict product and process specifications imposed by the government and health agencies for their production and commercialization. In addition to these constraints, companies are confronted with internal economical restrictions and market competition. Simulated Moving Bed (SMB) processes have proven to be an attractive and powerful separation technique [1] with great potential in this frame. Nevertheless, no unifying solution or approach for the automatization and control of SMB processes has managed to establish itself as a standard practice.
This thesis is based on the work of Erdem and Morari [2] which has proven to guarantee the fulfillment of product and process specifications, such as minimum purities and maximum allowable pressure drop, while optimizing the process economics during the unit operation. This approach merges the optimization and control problem, giving rise to a general constrained dynamic optimization based control concept.
The dynamics of the unit are described for control purposes by a linear time-varying reduced-order SMB model, that accounts for the mixed discrete/continuous nature of the process. This model is used on-line to predict and optimize the performance of the unit over a given prediction horizon. The four internal sectional ow rates are manipulated by acting on one internal and three external ow rates. Furthermore, Erdems approach makes use of a continuous monitoring system to get on-line measurements of the outlet concentrations and use them as the feedback information.
The development of the controller requires only the knowledge of basic information characterizing the system, i.e., the Henry's constants and average porosity of the columns. This is proven to be sufficient to control a SMB unit under linear and nonlinear chromatographic conditions. In this way, the time-consuming task of measuring nonlinear adsorption isotherms is reduced to the determination of the Henry's constants, which has a rather straightforward procedure. Moreover, the detailed characterization of each column is replaced by the evaluation of the average porosity of all the beds.
This thesis proposes an extension of the approach of Erdem and Morari [2], while retaining its core features and benefits. The novel property of the controller presented in this work, resides in its ability to make use of the concentration of the solutes at the outlets, averaged over one cycle, as the feedback information. This has special relevance, since measurement systems that collect samples over a given period of time, e.g., a cycle, are already in use as quality control systems in industry. Furthermore, the measurements of average concentrations, e.g., carried out in HPLC devices, are not affected by impurities or the presence of other species in multi-component systems, as in the case of online monitoring systems.
Here, the performance of the controller was thoroughly tested on a virtual SMB plant. Simulations were done on plants characterized by both, linear and nonlinear adsorption isotherms. The proposed control scheme proved to reject effectively the most common disturbances encountered in SMB practice. The controller was successfully implemented and tested under challenging scenarios on an lab-scale eight-column four-section SMB unit used for high purity separation of the nucleosides uridine and guanosine.

Supervisors: E. Gueltekin, Prof. Morbidelli, Prof. M. Mazzotti, Prof. M. Morari


Type of Publication:

(12)Diploma/Master Thesis

M. Morari

No Files for download available.
% Autogenerated BibTeX entry
@PhdThesis { Xxx:2005:IFA_2811
Permanent link