Incorporating feedforward action into self-optimising control policies



The available methods for the selection of controlled variables (CVs), using the concept of self-optimising control (SOC), focus on finding CVs, which are held at constant setpoints using a feedback controller. In this paper, it is shown that better self-optimising properties can be achieved by incorporating feedforward action using measured disturbances in the SOC policies. The resulting operational policy allows the setpoints for CVs to be varied with measured disturbances, allowing for near-optimal operation in the presence of disturbances and uncertainties. The effectiveness of the proposed approach is demonstrated through its application to an exothermic reactor and a forced-circulation evaporator.