Experimental study of rigorous nonlinear model predictive control for a packed distillation column

Dynamic modeling of distillation processes has been substantial in recent years. Our objective is to study the feasibility of real-time implementation of rigorous dynamic distillation models through on-line experimental studies. A rigorous model for a packed bed system is solved using the method of collocation polynomial. The control methodology used is a two-phase approach that consists of an on-line identification phase and an on-line optimization phase. The identification phase updates the key parameters of the rigorous model, while the optimization phase determines the control actions for the immediate future. The experimental system selected is a pilot-scale packed bed distillation process. The experimental studies show that the on-line control using this approach is feasible and performs better than traditional controllers.