(642i) Condition Based Maintenance for Sensor Network Reliability in Continuous Oral Solid Dose Manufacturing

The advent of Industry 4.0 (Smart Manufacturing) in the 2010s, is a manifestation of cultural and mindset change across multiple industries towards the acceptance of digital tools in process design and operations, resulting in highly automated and data-rich systems to realize lean manufacturing. These process intelligence tools leverage cross-industry advances in manufacturing methods, data analytics, fault management, and implementation technologies, for sustainable process operations, safety, quality risk assessment, and enhanced customer relations. In pharmaceutical systems, such digitalization tools facilitate the implementation of continuous manufacturing and real-time release testing by tracking product quality in real-time, thereby enabling the release of life critical medication with reduced off-line quality testing compared to existing practices.

Continuous manufacturing of drug product is the outcome of targeted process intensification with systematic integration of product and process knowledge, instrumentation and automation systems, quality control protocols and real-time process management for a solids processing system handling pharmaceutical materials. Success in the technology necessitates appropriate infrastructure for Information and Operations Technology integration.

At the core of such integrated systems is sensor network analytics, consisting of reliable and accurate measurements from field devices combined with mature process understanding to provide good quality estimates of unmeasured process variables [1]. In continuous tablet manufacturing, four leading causes of instrumentation faults are observed, which affect monitoring robustness. These include communication failure, instrument configuration, calibration and instrument age. Such instrumentation failures could potentially cause a systemic failure of the CM process.

This work focuses on leveraging systems engineering methods for real-time operations management, in particular, condition-based maintenance to improve operational reliability by mitigating sensor network failures in drug product CM. We discuss the following: (i) an integrated data-driven and model-based framework for systematic sensor network fault monitoring, (ii) the configuration and implementation of condition-based maintenance in a tablet processing pilot plant at Purdue University [2].

The proposed maintenance management approach is comprised of three workflows which are linked in an integrated data-driven and model-based framework. First, sensor network analytics methods of statistical process monitoring, data reconciliation, and gross error detection are leveraged for real-time process data validation, ensuring measurement accuracy, detecting sensor faults, and estimating unmeasured variables. Proof of concept of these approaches have recently been demonstrated by the research group [3,4]. Second, gross errors are identified for their root cause for appropriate maintenance activities, such as corrective, preventive and predictive maintenance. Third, to reconnect a maintained sensor, guidelines such as ASTM D6122, E2968 are leveraged and customized for the process.

The integrated framework is implemented at Purdue University in a pilot scale continuous direct compression facility. An analytically redundant sensor network using process equipment sensors, inline and at-line sensors are networked with automation systems such as Emerson DeltaV, OSIsoft PI System and Applied Materials Smart Factory Rx. A hierarchical systems integration architecture closely following ISA-95 is set up to advance the practice of Quality-by-Design in continuous manufacturing towards Quality-by-Control and Industry 4.0 [4].

References:

[1] Bagajewicz MJ. Smart Process Plants: Software and Hardware Solutions for Accurate Data and Profitable Operations. McGraw-Hill, 2009

[2] S. Ganesh, G. Reklaitis et al. Leveraging the digital infrastructure for real-time operations management in continuous pharmaceutical manufacturing: Condition-based maintenance for sensor network robustness (in preparation).

[3] M. Moreno, S. Ganesh, Y. Shah, Q. Su, M. Gonzalez, Z. Nagy, G. Reklaitis. Sensor Network Robustness using Model-based Data Reconciliation for Continuous Tablet Manufacturing. J Pharm Sci (2019).

[4] Q. Su, S. Ganesh, M. Moreno, Y. Bommireddy, M. Gonzalez, G. Reklaitis, Z. Nagy. A perspective on Quality-by-Control (QbC) in pharmaceutical continuous manufacturing. Comput Chem Eng (2019).