Theoretical and experimental study of polymer molecular weight distribution modeling and control in polymerization reactors

This dissertation describes a theoretical and experimental study of the control of molecular weight distribution (MWD) for linear polymers and copolymers produced by free radical addition polymerization. A new method is presented for modeling polymer molecular weight distribution. Equations are derived to calculate the weight fraction of a polymer's molecular weight distribution in different molecular weight, or chain length, regions. This approach leads to a simple method for choosing the overall molecular weight range to model as well as a potential method for modeling the tensile properties of certain amorphous polymers and a new method for modeling the molecular weight distribution of copolymers. This is a departure from previous methods which determine the weight fraction of polymer only at specific values of molecular weight.;Experiments involving the free radical solution polymerisation of methyl methacrylate in a batch reactor are done to demonstrate the use of this MWD modeling approach. Furthermore, a new method is developed to control the MWD of the polymer product on-line. Control of the MWD is achieved by computing a suboptimal sequence of reactor temperature setpoints which lead to the best match of a given target MWD at the end of the batch. This sequence is solved using sequential quadratic programming. An on-line extended Kalman filter is used to incorporate infrequent off-line molecular weight distribution measurements.