Lou Heavener, Emerson and Markus Krumme, Novartis
The pharmaceutical industry is currently based on batch processing predominantly and moving to adopt continuous processing as next generation of a manufacturing architecture. There are obvious economic benefits, but also important concerns relating to meeting validation and regulatory requirements that need to be overcome. For the purpose of this paper, continuous processes would be those with a constant product draw, which introduces new challenges to quality assurance. Fortunately, other industries have resolved some of these issues and the petroleum refining industry is considered as a case study for continuous control. They have variable feedstocks and complex product specifications. They face seasonal demand swings and must be prepared to handle supply problems, challenged at the same time with variable incoming raw material qualities and cost pressure. From an automation and control perspective, on-line analyzers, inferential sensors, and advanced control technologies such as model-predictive-control have been employed to overcome challenges and demonstrate adaptive control techniques as enabler of maintaining strict but tunable quality specifications for the end products, being constrained by variable incoming raw material qualities.
Key message in the talk is the demonstration of a different approach to assure unadulterated quality products by a more flexible control structure and different regulatory constraints in a highly cost-sensitive industry and trigger learnings which we can adopt from there into the pharmaceutical world.