Nonlinear model predictive control with state estimation in batch polymerization

A computationally simple and practical nonlinear control algorithm is developed for temperature control of a batch styrene polymerization reactor. Calculation of control effort is accomplished by a Newton-type single step strategy, combined with a state estimation algorithm. To realize the state estimation, the rigorous system composed of six state variables is divided into subsystems and two extended Kalman filters (EKF) with reduced orders are used for every subsystem. The control and estimation algorithm is implemented in an interactive real-time data logging, monitoring and control environment through a graphical user interface. Simulation runs revealed good set point tracking during the isothermal stage of the batch operation.