(623d) A Fault-Tolerant Control Design for Real-Time Release in Continuous Manufacturing of Solid Dose Using Direct Compaction

Qinglin Su, Mariana Moreno, Jianfeng Liu, Sudarshan Ganesh, Yasasvi Bommiready, Marcial Gonzalez, Gintaras V. Reklaitis, Zoltan K. Nagy

School of Chemical Engineering, Purdue University,
West Lafayette, IN 47907

Thomas O’Connor, Tian Geng

CDER, Office of Pharmaceutical Quality, U.S. Food and Drug Administration,
Silver Spring, MD 20993

Abstract

Advantages of continuous manufacturing in the pharmaceutical industry have been demonstrated through extensive research studies in the last decade [1], but mostly at the conceptual or theoretical level. Only few continuous manufacturing processes have been approved by the US Food and Drug Administration thus far, e.g., the Vertex Orkambi and Janssen Darunavir [2]. To move from a conceptual continuous design to a practically operating pilot plant or manufacturing process, the efforts of regulators, academics, and industry are now focused more on quality control strategies [3]. Specifically, the active process control design for real-time release and its risk management in continuous manufacturing [4, 5] have been considered as critical for successful implementation of continuous manufacturing processes.

In this study, a fault-tolerant control design for real-time release is implemented, according to our previous systematic framework for control design and analysis [5], to a pilot plant at Purdue University for continuous manufacturing of solid dosage pharmaceuticals using direct compaction. First, beside the fundamental control loops provided by the equipment vendors for unit operations of feeding, blending, tableting, etc., to manipulate the critical processing parameters (CPPs), supervisory active feedback and feedforward process control strategies are designed to maintain the critical quality attributes (CQAs) based on process analytical technology (PAT) tools and final product quality testing, e.g., Sotax Auto Test 4. In the active process control design, the major system dynamics are identified with state-space models using historical operating data in the pilot plant, and a series of rational control design and analysis metrics are employed to evaluate the control performance and screen the potential design risks [5]. For example, a Relative Gain Array (RGA) analysis is used to evaluate the system interactions and select the best-performing control pairings in the feeding-blending units, where the diagonal pairing is preferred under low system frequency domain, and is risky under higher frequency as is the case when the feeder nozzle is accidently attached to the blender hopper frame.

In the proposed fault-tolerant control design, a risk mapping, assessment and planning (Risk MAP) strategy is also considered to address the potential risks associated with the above active process control design. The active process control system is designed to mitigate or avoid quality risks. However, implementation of risk reduction measures can introduce new risks into the system or increase the significance of other existing risks (ICH Q9). In this study accepted residual risks and new potential risks to product quality after the implementation of the active process control system for the continuous manufacturing process are examined. . For example, when faults occurred in the feeding-blending units, the process performance indices, e.g., process capabilities index, are below their minimum targets. Hence, for real-time release purposes, the proposed fault-tolerant control design is evaluated experimentally and in simulation mode under these risk scenarios, e.g., process disturbance, probe fouling, instrument calibration errors, etc. It is as expected that the fault-tolerant control design with advanced process systems engineering tools, such as process monitoring, data reconciliation, state estimation, etc., can mitigate those potential risks that are common causes of variations in an otherwise stable process.

Steady-state operation of the pilot-plant is also executed over longer durations to validate the proposed fault-tolerant control design. Concluding remark will also be offered, suggesting potential guidelines for achieving real-time release in continuous solid dosage manufacturing.

References