Closed-loop Implementation of Optimal Trajectories in Batch Distillation

HECTOR JOSE MARIA ESPINOSA; LUIS MARCHETTI, JACINTO

Bucharest

Simposio; 17th European Symposium on Computer Aided Process Engineering; 2007

Fractionation in batch distillation columns is one of the most common technologies used in pharmaceutical and specialty chemical industries due to its operational flexibility. A batch column can separate multicomponent mixtures with a wide range of feed compositions and different degrees of separation difficulty. During the last decade, several authors worked in the development of methods to cope with different steps of the design and synthesis of batch distillation systems. Among these methods, conceptual models based on pinch theory have been successfully applied to determine quasi-optimal trajectories intended to obtain products with both purities and recoveries above a certain level while operating the column near the condition of minimum energy demand [1, 2]. In the case of batch reactive distillation, the issue is maximum feasible product recovery by distillation to enhance both limiting reagent conversion and reaction rate in the still [3, 4]. Once the quasi-optimal operation is determined, the practical implementation requires an appropriate control system adjusted to track the desired conditions while maintaining the light-component purity free from unfavorable disturbances. This work proposes a temperature tracking control system composed by an open-loop reflux ratio control plus a closed-loop correction for disturbance rejection. To adjust the closed-loop controller we use the referential dynamic reaction of the process and tuning rules [5] that, though they were originally developed for dynamics valid in the neighborhood of stationary operating points, under this strategy they are useful in the neighborhood of a reference transient evolution like those occurring in batch distillation columns. Previous successful experiences in tracking a desired temperature evolution of a bioreactor [6] motivate the present application. Here, binary and ternary separations of alcoholic mixtures are selected as case studies. In the first section, the conceptual model based on pinch theory is presented for both binary and multicomponent mixtures whilst the evolution of the reflux ratio necessary to obtain high purity distillate composition is reported for the nominal case of an equimolar binary mixture of alcohols. After that, the structure of the control system is presented, and a tray temperature is selected as controlled variable based on its sensitivity to differences in the initial composition of the reboiler content. Then, the referential reaction curve is obtained by introducing changes into the nominal trajectory of the distillate flow rate. This reaction curve allows the estimation of parameters of a model useful for tuning the feedback controller. Results of rigorous closed-loop simulations by changing the initial still composition are presented. Finally, the performance of the proposed methodology to track optimal conditions for a ternary mixture of alcohols and the conclusions are presented together with comments about our future work. Keywords: Batch Distillation, Conceptual Model, Process Reaction Method. Related Literature [1] Espinosa, J.; Salomone, E.; Iribarren, O. Computer-Aided Conceptual Design of Batch Distillation Systems. Ind. Eng. Chem. Res. 2004, 43, 1723. [2] Brüggemann, S.; Oldenburg, J.; Marquardt, W. Combining Conceptual and Detailed Methods for Batch Distillation Process Design. In: FOCAPO 2004, 247. [3] Sorensen, E.; Macchietto, S.; Stuart, G.; Skogestad, S. Optimal Control and On-Line Operation of Reactive Batch Distillation. Computers chem. Engng. 1996, 20 (12), 1491. [4] Espinosa, J. Assessing the Performance of Batch Reactive Distillations through Conceptual Models. In: ESCAPE-16 2006, 433. [5] Ziegler, J. G.; Nichols, N. B. Optimum Settings for Automatic Controllers. Trans. ASME, 1942, 64, 759. [6] Marchetti, J. L. Referential Process-Reaction Curve for Batch Operations. AIChE J. 2004, 50, 3161.