Development of on-Line Optimization-Based Control Strategies for a Starved-Feed Semi-Batch Copolymerization Reactor

A real-time nonlinear optimization scheme is implemented for the high temperature starved feed semi-batch solution polymerization used to produce acrylic resins for automotive coatings system. The system input and state trajectories are determined on-line to minimize a cost function defined in terms of deviation from target molecular weight and composition, and batch time. The performance of the proposed strategy is examined through numerical simulation for several study cases for semibatch butyl methacrylate and styrene (BMA/STY) free radical copolymerization. Results show that polymer quality can be assured in reduced batch time while achieving constraint satisfaction over the entire run. Efforts to apply this control scheme to an experimental reactor system will also be discussed.