(778f) Advanced Model Predictive Control of Powder Level in Continuous Pharmaceutical Manufacturing Pilot-Plant

Authors: 
Singh, R., Rutgers, The State University of New Jersey
Muzzio, F., Rutgers, The State University of New Jersey
Ierapetritou, M., Rutgers, The State University of New Jersey
Ramachandran, R., Rutgers University
Currently, pharmaceutical companies are going under paradigm shift from batch to continuous manufacturing. In continuous pharmaceutical manufacturing (CPM), real-time effective control of Critical Process Parameters (CPPs) and Critical Quality Attributes (CQAs) is necessary to consistently produce a very high quality product as mandated by regulatory authority and needed for drug efficiency and patient safety [1]. Efficient automated coordination among different unit operations involved in continuous manufacturing is prerequisite to achieve consistent flow rate, inlet composition, line holdup, and thereby consistent product quality using QbD principles. In CPM, blender is connected with tablet press through a transfer pipe (chute). The chute also provides a suitable interface to integrate sensors (e.g. NIR, Raman) for real time monitoring and control. Maintaining consistent powder level in chute is needed to assure the presence of powder in front of sensors, to avoid over flow and under flow of the powder and to achieve the consistent product quality. However, different level of disturbances, change in raw materials and product specifications and startup, shutdown, speedup, slowdown processes could affect the powder level in chute significantly and thereby process operation, monitoring and product quality. Therefore, real time monitoring and control of powder level is necessary for continuous pharmaceutical manufacturing.

In this work, an advanced model predictive control strategy as well as a PID based control strategy for powder level control in a chute placed in between blender and tablet press unit operation of continuous tablet manufacturing process has been developed, implemented and evaluated. This work has been accomplished in three phases. In first phase, the Insilco study employing the process and sensor model has been performed to identify the best sensing and control strategy for powder level. Two types of the sensors, the one that gives discrete signal (on/off) and the one that gives continuous signal can be used for real time monitoring of powder level in chute. Discrete and continuous control strategy has been developed specifically for these two types of the sensors. The performance of two types of the sensor for real time monitoring and control of powder level has been compared Insilco. The performance of advanced model predictive controller (MPC) for powder level has been also compared with PID controller. We found that a continuous control strategy is more effective than a discrete control strategy, and that a MPC control strategy is more effective than a PID control strategy. In second phase, a novel noninvasive technique based on change in electric field concept has been developed, integrated and evaluated for real time monitoring of powder level in continuous manufacturing pilot-plant. The sensor has been integrated with a control panel (DeltaV (Emerson)) through relays and charms. The control panel then send the signal (4-20 mA) to control platform where the control strategy, a program to operate the sensor and a program to convert the sensor signal into powder level have been implemented. The performance of electric field based sensor for real time monitoring of powder level has been compared with a webcam based sensor. In third phase, an advanced model predictive controller as well as PID controller have been implemented into our continuous pharmaceutical manufacturing pilot-plant facility. The performance of both control strategies have been practically evaluated for set point tracking and disturbance rejection via utilizing electric filed based sensor. The MPC performs better than PID.

The objective of this presentation is two-fold; first to highlight the Insilco studies for identification of best sensing and control strategy for power level and then demonstrate the performance of best sensing and control strategies into our continuous pharmaceutical manufacturing pilot-plant facility

Reference

  1. Singh, R., Román-Ospino, A. D., Romañach, R. J., Ierapetritou, M., Ramachandran, R. (2015). Real time monitoring of powder blend bulk density for coupled feed-forward/feed-back control of a continuous direct compaction tablet manufacturing process. International Journal of Pharmaceutics, 495, 612-625. http://dx.doi.org/10.1016/j.ijpharm.2015.09.029