Estimation and Control of Continuous Stirred Tank Polymerization Reactors

The extended Kalman filter has been applied to a continuous styrene polymerization process with mixed initiators for the on-line state estimation and control. Two time scale filter is used with the reactor model having some inherent modeling error. The effects of measured and unmeasured disturbances, and the effect of long off-line measurement delay have been investigated. It is also shown that the filter can be used successfully for rapid polymer grade changes.