471687 Feedforward Control of Continuous Pharmaceutical Manufacturing Process

Friday, November 18, 2016: 8:30 AM
Continental 4 (Hilton San Francisco Union Square)
Ravendra Singh, Glinka Cathy Pereira, Nikita Soni, Andrés D. Román-Ospino, Marianthi Ierapetritou and Rohit Ramachandran, Department of Chemical and Biochemical Engineering, Rutgers, The State University of New Jersey, Piscataway, NJ

Continuous manufacturing has been in use in many industries for a long time, however the interest in pharmaceutical continuous manufacturing is relatively new and most pharmaceutical companies are now moving from batch to continuous. With continuous manufacturing we have shorter processing times, increased efficiency and increased product quality assurance in real time. The product quality is of utmost importance in pharmaceutical manufacturing and in order to control that, we need to make sure that effect of input disturbances are proactively mitigated before they affect end product quality. Therefore it is desired that a combined feedforward/feedback control system be implemented over a traditional feedback control system to dissipate and control the disturbances from entering the process. The feedforward controller measures and takes corrective actions for the disturbances before they affect the process while the feedback controller ensures the consistency of the output [1].

In this work, feedforward control loops have been designed for continuous pharmaceutical manufacturing process via wet granulation (WG), roller compaction (RC) and direct compaction (DC) and their performance have been evaluated through a flexible integrated process flowsheet model. For WG line, a combined feedforward/feedback control system has been developed to control the drug concentration at twin screw granulator (TSG) outlet. This process involves a batch pre-blending step where API and intragranular components are mixed, and subsequently fed into a continuous wet granulation (TSG) processing step, together with second stream containing excipients. The commingling of these two streams in the twin screw granulator leads to an overall low API concentration. This concentration of the API is very small and hence, difficult to be measured by a Near Infrared (NIR) sensor. Thus, the role of the residence time distribution (RTD) model in predicting the concentration at the outlet of the TSG as well as to determine the disturbance dissipation along the process becomes significant. Along with this, RTD can also be used for the raw material traceability and scale-up requirements. A major source of drug concentration variability is the potential lack of homogeneity of the stream entering the granulator. Pre-blends discharged by a tumbling blender are known to fluctuate fairly significantly in composition and due to limited back mixing in the twin screw granulator, these fluctuations are propagated to the granulator exit, affecting the content uniformity of API in the granules. In turn, such variability can affect product content uniformity (CU). Since the composition of the stream discharged by the feeder can be monitored instantaneously using PAT method, it is therefore possible to adjust the ratio of the API-bearing stream and the excipient-only stream, and their flow rates, to compensate for composition fluctuations entering, and therefore exiting, the granulator. This has been achieved through a feedforward control mechanism combined with a feedback control loop. For RC line, a feed forward control loop has been developed to mitigate the effect of lubricant concentration. This loop has been integrated with feedback control loop for ribbon density which is cascade with hydraulic pressure feedback (slave) control loop, and a feedback control loop for throughput. For DC line, three feedforward control loops have been developed to proactively mitigate the effects of variations of powder bulk density, lubricant concentration and drug concentration on tablet weight & hardness, tablet hardness, and tablet potency respectively. However, maximum two feed forward control loops can be activated at a time. These feedforward loops have been integrated with feedback control loops [2].

The aim of this presentation is two-fold. First to highlight the feedforward control loops involved in continuous pharmaceutical manufacturing via WG, RC and DC and second to demonstrate the performance of combined feedforward/feedback control system over only feedback control system.


[1]. Singh, R., Muzzio, F., Ierapetritou, M., Ramachandran, R. (2015). A combined feed-forward/feed-back control system for a QbD based continuous tablet manufacturing process. PROCESSES Journal, 3, 339-356.

[2]. Singh, R., Sahay, A., Karry, K. M., Muzzio, F., Ierapetritou, M., Ramachandran, R. (2014). Implementation of a hybrid MPC-PID control strategy using PAT tools into a direct compaction continuous pharmaceutical tablet manufacturing pilot-plant. International Journal of Pharmaceutics, 473, 38–54.

Extended Abstract: File Not Uploaded