Experimental study of optimizing control of continuous chromatographic separation process

This thesis introduced an improved single-column chromatographic (ISCC) separation process with the final objective to make this process distinct from existing single-column chromatographic separation processes by physical modifications and conceptual advances. The performance of this ISCC process was evaluated by experimental implementation to separate a mixture of guaifenesin enantiomers. In ISCC process, different standard HPLC peripherals were used as building blocks and some standard parts of the commercial HPLC system were redesigned to overcome the existing limitations for better performance. Fraction collection schemes and mechanism are the important features of this improvement. This fraction collection system allows accommodation of overlapped peaks from adjacent cycles and reduce the overall time delay of the process. These process design modification provide a wider degree of freedom: injection volume, cycle time, desorbent flow rate, feed concentration and fraction collection intervals. A robust online monitoring system was designed which was relatively inexpensive and was able to offer high frequency and accurate analysis of the samples compared to other devices. The proposed ISCC process was assembled in a laboratory and commissioned successfully. Process performance was optimized by a multi-objective stochastic optimization technique based on genetic algorithm (GA). The optimization problem was appropriately formulated with the aim of maximization of productivity and minimization of desorbent requirement. Performance of the ISCC process was also compared with a similar SMB process. This study provided the basis for reaping the full potential benefits of a single column process that adopts cyclic injection. Besides, relative contribution of the decision variables were ascertained through the study of their effects on the performance indicators. Detector calibration and determination of adsorption isotherm parameters were done simultaneously by adopting a new method named nonlinear direct inverse method, which is relatively fast, and economical technique compared to existing alternatives. A `cycle to cycle’ model predictive control (MPC) scheme was developed in-house to guarantee product and process specifications for obtaining optimized profitability. The performance of this MPC scheme was demonstrated through simulation studies. Finally, the cycle-to-cycle optimizing controller developed for the ISCC process for the separation of a mixture of guaifenesin enantiomers. Key implementation issues were accuracy of the online measurement system and integration and automation of the ISCC process with online measurement system and controller. This was achieved by designing and developing a human machine interface (HMI) that was able to effectively communicate among the three essential components of the control loop. The performance of the controller was tested for set point tracking and disturbance rejection. Results indicate that the designed ISCC process with the online monitoring system was able to run at the optimal operating conditions and deliver the product requirements as confirmed by open-loop and close-loop experiments.DOCTOR OF PHILOSOPHY (SCBE