(586o) Design of An Efficient Control System for Flexible Continuous Tablet Manufacturing Process

Monitoring and close loop control is of great important in continuous manufacturing of pharmaceutical tablet to guarantee the end product quality. In the work reported here, an efficient control system has been designed for a flexible multipurpose continuous tablet manufacturing process. The flexible continuous tablet manufacturing process consists of direct compaction (DC) [1], wet granulation (WG) and dry granulation (DG) [2] routes where one route can be selected based on feed powder characteristics and the desired end product quality specifications. DC is the simplest route used for easily flowing material, the WG route can be used to improve the product quality, and the DG route can be used in place of WG if the API and/or Excipient are water sensitive. Three control strategies (PID, Model Predictive Control (MPC), and, hybrid MPC-PID) have been designed for this process and their performance regarding set point tracking and disturbance rejection has been compared using a flowsheet model simulated in gPROMS (PSE) and control interface DeltaV (Emersion).

The control systems have been implemented in the control studio of the DeltaV system and the inputs and outputs of the controller are connected with the process model simulated in gPROMS. IGEAR software and the gORUN feature of the gPROMS have been used to communicate between gPROMS and DeltaV system. The model is used in place of the plant to generate the data needed for design and performance evaluation of the control system. In direct compaction process the API composition, Relative Standard Deviation (RSD) and flow rate are controlled at blender, and the tablet weight and hardness are controlled at the tablet press. The selected control strategy is then integrated with the pilot plant through control hardware, software and sensors implementation.

NIR is used to monitor the API composition at blender outlet, CAMO Unscrambler X is used to develop the prediction model, Unscrambler Process Pulse uses Prediction engine and the prediction model to generate the composition signal from the measured spectrum. This signal is then communicated to the DeltaV control studio through MATLAB OPC communication protocol. Based on the API composition measurement, the controller implemented in DeltaV calculates the actuator setting (feeders rotation speeds) which is send to the plant through DeviceNet. Similarly, RSD a measure of blend uniformity is controlled using NIR as the monitoring tool and blender rotational speed as the actuator. The tablet weight is controlled through a cascade control scheme where the master controller generates the set point of pre-compression force, which is then tracked by a slave controller by manipulating the lower punch displacement of the tablet press. The lower punch displacement changes the amount of powder filled in the die. The tablet hardness is also controlled through a cascade control scheme where the master controller generates the set point for main compression force which is tracked by a slave controller by manipulating the upper punch displacement. The pre-compression and the main-compression forces are monitored through strain gauge (sensors inbuilt in the tablet press) while the tablet weight is monitored through a load cell/weighing instrument and NIR is used for hardness measurement. Alternatively, soft sensing approach can be also employed for weight and hardness measurements.  

The objective of this presentation is two-fold: first to demonstrate the model-based design and performance evaluation of the different control strategies for flexible multipurpose continuous tablet manufacturing process and then to present the implementation of the best control strategy into the direct compaction line.      

References

1.  Singh, R., Ierapetritou, M., Ramachandran, R. (2013). European Journal of Pharmaceutics and Biopharmaceutics, http://dx.doi.org/10.1016/j.ejpb.2013.02.019.

2.  Singh, R., Ierapetritou, M., Ramachandran, R. (2012). International Journal of Pharmaceutics, 438 (1-2), 307-326.