REAL-TIME CONTROL AND OPTIMIZATION OF BATCH FREE-RADICAL COPOLYMERIZATION REACTORS

An exogenous-variable self-tuning controller is developed which incorporates separate proportional and integral action in the cost functional. Guidelines are given in order to implement this self-tuner effectively in a real-time environment.; A general, minimal-order structure for batch reactors which display moderate exothermicity is presented. Simulation results show that this structure, when used in conjunction with the self-tuner, yields good temperature-tracking performance for copolymerization reactors.; Rejection of large exothermic loads in batch reactors can be achieved by using a cascade controller, which is composed of a self-tuner and a classical (proportional-integral-derivative) controller. This cascade controller can significantly improve the dynamic response of systems which employ poorly tuned classical control algorithms without recourse to re-tuning.; An adaptive optimizer is derived which can be employed to improve copolymer properties such as copolymer composition distribution, sequence distribution, and polydispersity ratio. The algorithm requires only one forward integration of the copolymer system equations. Furthermore, it is demonstrated by means of simulation studies that this algorithm can employ realistic sampling times to achieve optimization of copolymer properties in the presence of severe measurement noise, and is therefore implementable in real-time.; Experiments that employed the free-radical solution copolymerization of styrene-methylmethacrylate were carried out in a one-gallon batch reactor. The results from these experiments demonstrate the feasibility of the "off-line" adaptive optimizer as well as the self-tuning controller. Discrepancies between the copolymer model and the observed data indicate a need for "on-line" optimization and/or improvements in the model parameters.