Self-adaptive predictive functional control of the temperature in an exothermic batch reactor

In this paper we study a self-adaptive predictive functional control algorithm as an approach to the control of the temperature in an exothermic batch reactor. The batch reactor is located in a pharmaceutical company in Slovenia and is used in the production of medicines. Due to mixed discrete and continuous inputs the reactor is considered as a hybrid system. The model of the reactor used for the simulation experiment is explained in the paper. Next, we assumed an exothermic chemical reaction that is carried out in the reactor core. The dynamics of the chemical reaction that comply with the Arrhenius relation have been well documented in the literature and are also summarized in the paper. In addition, the online recursive least-squares identification of the process parameters and the self-adaptive predictive functional control algorithm are thoroughly explained. We tested the proposed approach on the batch-reactor simulation example that included the exothermic chemical reaction kinetic model. The results suggest that such an implementation meets the control demands, despite the strongly exothermic nature of the chemical reaction. The reference is suitably tracked, which results in a shorter overall batch-time. In addition, there is no overshoot of the controlled variable T, which yields a higher-quality production. Finally, by introducing a suitable discrete switching logic in order to deal with the hybrid nature of the batch reactor, we were able to reduce the switching of the on/off valves to a minimum and therefore relieve the wear-out of the actuators as well as reduce the energy consumption needed for control.