A Model Predictive Control Scheme for Safe and Optimal Operation of Exothermic Semi-Batch Reactors

Abstract In this work a model predictive control scheme for on-line optimization of strongly exothermic semi-batch reactors is presented. In addition to the usual operational constraints, such as those on the manipulated variables, safety requirements are explicitly incorporated into the problem formulation even for the case of possible failure scenarios. The proposed method determines optimal trajectories for the existing manipulated variables in the prediction horizon while analyzing the process behavior within certain time windows specific for each failure to be considered. In order to enhance the tractability of the resulting complex optimization problem several simplifications are discussed. The derived concept is illustrated through simulation studies of a polymerization reactor.