Modeling, process monitoring and optimizing control of emulsion polymerization processes

Emulsion polymerization is an important industrial (semi-) batch process for the production of latices, which are used as adhesives and paints. On the growing, but very competitive polymer market, a continuous process improvement is essential. work. Basis is a simulation model for a radical emulsion polymerization, which has been calibrated to the performance of styrene and butyl acrylate homo- and co-polymerizations. Using that model, the process observability and state estimation of emulsion polymerization processes with temper- ature measurements is analyzed and the performance of different Extended Kalman Filters is examined in presence of plant-model mismatches. The results show, that temperature measurements alone do not enable a robust estimation of the inner process states in a co-polymerization process. Therefore, the use of an additional ultrasound velocity measurements is discussed, motivated by the difference in the sound velocity of water, monomers, and polymer. As all available literature approaches to describe the sound velocity as a function of the mixture composition have their limitations, an own model is derived. Afterwards, the impact of temperature and sound velocity fusion on the estimator performance is examined, illustrating that the monitoring robustness versus kinetic parameter mismatches is improved. At last, different control schemes are tested for batch time minimization and production of desired co-polymer compositions. It is demonstrated that batch time reduction does not require model-based control, while a model-predictive control scheme to adjust the polymer properties does not improve the results of a classical PI control approach in the applied example.