(444c) A Plant-Wide Control Strategy for Continuous Pharmaceutical Manufacturing

The pharmaceutical industry is historically dominated by batch-wise processing. However, the industry is challenged by demands for faster time-to-market, tighter control over product quality, and a smaller ecological foot print. Batch processes suffer from several inherent drawbacks such as poor scale-up behavior and inhomogeneous process conditions [1]. A step change is needed to improve the performance of current pharmaceutical processes drastically. Continuous manufacturing has the potential to provide such a step change. Continuous processes are simpler in nature as the time dependency of process variables vanishes at steady state operation. Tight operation around a desired steady state can reduce the amount of off-spec product, which remains a notorious challenge in batch-wise processing. An effective control system is of vital importance to enable continuous pharmaceutical processes. Quality by design can be assured by a control system that is based on fundamental process understanding. In this contribution, a plant-wide control strategy is presented for a pilot plant continuous pharmaceutical process. The pilot plant is being constructed within the Novartis-MIT Center for Continuous Manufacturing at MIT and produces an active pharmaceutical ingredient from start (synthesis) to finish (tablet formation) in a continuous fashion. A systems view is adopted to optimize the performance of the process as a whole instead of optimizing the individual unit operations. A hierarchical decomposition strategy is used to deal with the complexity of the problem [2] as the control system design needs to address several implicit and explicit control objectives by fixing a large number of (discrete) decision variables. At the highest decomposition level, control objectives related to the inputs and outputs of the process are optimized. The objectives are refined and new control objectives are added upon increasing the level of detail of the process representation. The most detailed level involves individual unit operations. Note that this decomposition strategy naturally classifies the characteristic time constants of the process disturbances and exogenous inputs from slow acting at the input-output level (e.g. steady state transitions) to fast acting at the level of individual unit operations (e.g. temperature disturbances). A basic process model is developed to evaluate the performance of different control structures. At the highest decomposition level, steady state models are sufficient to capture the response of the system towards disturbances and exogenous inputs. For the faster processes, at the lower decomposition levels, dynamic models are used to evaluate the propagation of disturbances within the process. In addition, effective control systems need to be designed for novel processing technologies such as microfluidic devices [3]. The studied pilot plant contains several implementations of microfluidic devices. Continuous pharmaceutical processes are excellent candidates due to the typically small production capacities, often hazardous chemical systems, and high added-value products. The precise control of flow rates and pressure drops for microfluidic devices is still challenging and important for successful operation in a production environment [4]. An effective control system is a crucial part of future continuous pharmaceutical processes. This work provides a case study of how to use a hierarchical decomposition strategy to deal with the complexity of controlling these novel processes. Future research needs include the development and implementation of more advanced control strategies such as model predictive control, which would further enhance the flexibility and robustness towards process variability.

References

[1] Plumb, K. Continuous processing in the pharmaceutical industry: Changing the mind set. Chem. Eng. Res. Des., 2005, 83, 730?738.

[2] Stephanopoulos, G.; Ng, C. Perspectives on the synthesis of plant-wide control structures, J. Process Control, 2000, 10, 97-111.

[3] Jensen, K.F. Microreaction engineering - Is small better? Chem.Eng.Sci., 2001, 56,293-303.

[4] Hartman, R.; Jensen, K.F. Microchemical systems for continuous-flow synthesis. Lab Chip, 2009, 9, 2495-2507. .