The manufacture of pharmaceutical products has in recent times started shifting from a batch processing paradigm to a continuous manufacturing one. Motivations for this shift include lower cost and time of manufacturing for drug products, the availability of new online sensing and instrumentation technologies, and competition from generic drug product manufacturers. Continuous manufacturing of pharmaceutical products also entails certain challenges in real-time process management. One set of challenges arises from the use of bulk powders, gels, suspensions, and other media that show non-ideal behaviors. Processing such media requires novel approaches in process control techniques. A second set of challenges arise in integrating and managing a very large amount of knowledge about process conditions and product formulations, which is needed to flexibly and reliably scale processes and transfer knowhow. The nature and structure of such knowledge makes it ill-suited for DBMS systems, and required novel techniques in informatics.
In this work, we present results and case studies from the Engineering Research Center for Structure Organic Particulate Systems (ERC-SOPS), a collaborative effort between academia and industry in improving pharmaceutical manufacturing. We show our advances in real-time process management, including model predictive control, fault detection & diagnosis, real-time optimization, as well as the knowledge management techniques that enable the rapid deployment and scaleup of process management strategies. We present case studies in the manufacture of various dosage forms, focusing especially on tablets and drop-on-demand products. Finally we summarize the challenges still ahead and offer general conclusions for the operations and systems communities in enabling continuous manufacture for pharmaceutical products.