Evolutionary design, multiobjective optimization and control of chromatographic processes

Although preparative ion-exchange chromatography (IEC) is widely employed for protein purification, the choice of operating conditions has remained largely empirical, resulting in sub-optimal performance of these separation systems. The lack of an appropriate theoretical framework for nonlinear protein chromatography has limited the development of efficient nonlinear operations. Recent advances in the theory of protein nonlinear ion exchange chromatography along with the development of novel low molecular weight displacers set the stage for the development of optimal gradient and displacement purification processes. A novel hybrid model framework that provides flexibility to the chromatographic engineer and dramatically reduces the computational time required for simulation and multivariable optimization of preparative chromatographic processes is presented. It also enables the estimation of optimal operating conditions, under different parametric specifications without any additional computational requirements. In addition, a physical programming based framework for multiobjective optimization is developed for chromatographic processes. We have demonstrated that this strategy is well suited to address the priorities and tradeoffs of various competing objectives and/or constraints in complex non-linear chromatographic systems. This approach provides flexibility to the chromatographic engineer and enables the use of physically significant ranges of desirability (e.g., desirable, tolerable, undesirable, etc.) for each design objective. In order to improve the performance of large-scale chromatographic processes by reducing batch-to-batch variations we have developed a novel Generalized Run-to-Run Control (GR2R) technique that has ability for simultaneous optimization and control of non-linear chromatographic processes. GR2R strategy can optimize and control the process in the presence of batch to batch and sporadic variations. The results presented demonstrate that the current practice of using fixed optimal conditions can lead to sub-optimal performance in the presence of measured and unmeasured disturbances. In contrast, processes using generalized run to run control can result in an improved performance in the presence of a variety of disturbances. The presented GR2R strategy is highly efficient in rejecting disturbances as well as for target tracking.