STARTUP OF A DISTILLATION COLUMN USING NONLINEAR ANALYTICAL MODEL-PREDICTIVE CONTROL

This paper discusses the application of nonlinear analytical model predictive control to distillation column startup between the time when the trays are hydraulically sealed and the time steady-state operation is reached. A novel startup method is proposed whereby the liquid feed is initially used as reflux to seal the trays by establishing the plate holdups. The control algorithm utilizes a nonlinear analytical process model to generate an optimal control vector trajectory, which is implemented by cascade controllers. The analytical model, based on material and energy balances, additionally utilizes the secondary process measurements resulting in improved controller performance. As the steady state is approached, a neighboring optimal control signal is superimposed on the output from the supervisory level controller to eliminate fast disturbances entering the column. In the hierarchical control structure, the adaptive online optimizer resides in the PC/AT which is networked with PC/XTs for neighboring optimal and cascade controls. A PID type controller with periodic update of set points and a manual startup have been compared with the present algorithm. Experimental results showed a significant reduction in time as well as cost in the case of the proposed scheme.