POD-based predictive controller with temperature constraints for a tubular reactor

In this paper the control of a non-isothermal tubular reactor is addressed using proper orthogonal decomposition (POD) and predictive control techniques. The proposed control system has the goal of keeping the operation of the reactor around a given operating condition in spite of the disturbances in the feeding flow. Additionally the controller has to avoid that the temperature inside the reactor be larger than a given value in order to prevent undesirable side reactions. The proposed control system handles this constraint as a soft-constraint and it uses a slack variable approach with linfin- norm and time-dependent weights to handle the infeasibilities that can emerge. The MPC controller is based on a model derived by means of the POD technique, which reduces the high dimensionality of the discretized system used to approximate the partial differential equations (PDEs) that model the reactor. The simulation results show that the predictive controller presented a good behavior and that it dealt with the temperature constraint quite well.