MODELING, MONITORING, AND CONTROL OF PH-SHIFT REACTIVE CRYSTALLIZATION

Abstract

Crystallization is one of the most important unit operations for separation and purification in process manufacturing industry. With the recent research development prompted by the process analytical techniques (PAT) in batch cooling and antisolvent crystallization processes, the reactive crystallization process has also been receiving increasing interest due to its importance and wide application in industry. This thesis first developed a first-principles model for a semi-batch pH-shift reactive crystallization process using the polymorphic L-glutamic acid as a model compound. Next, a moving-window multiway principle components analysis (MPCA) was presented for process monitoring. Direct design control strategies based on tracking nominal optimal concentration and polymorphic purity trajectories were discussed and compared. Furthermore, nonlinear model predictive control based on Extended Prediction Self-Adaptive Control (EPSAC) technique using state-space models formulation, as well as its integration with Batch-to-Batch (B2B) control, viz., B2B-NMPC, were also investigated under model-plant mismatches.