AI digital twin developed to boost pharma manufacturing efficiency

With a digital twin, virtual modelling is combined with monitoring, analysis and simulation of a physical system

A new AI digital twin platform aims to enhance fault detection, system monitoring and predictive maintenance in order to boost pharma manufacturing efficiency and reliability.

The newly-developed system uses artificial intelligence (AI) and real-time plant data to automate the development of digital twins that act as virtual replicas of production lines. These can then be used to optimise plant operations, detect anomalies such as mismatched flow rates or abnormal tank levels before they escalate.

The technology is the result of a collaboration between University of Cambridge overseas research centre – Cambridge Advanced Research and Education in Singapore (CARES) – and the A*STAR Institute for Infocomm Research (A*STAR I2R), which is part of Singapore’s public sector R&D agency.

The partners mapped data to connected ontologies, which present knowledge in a structured, machine-readable way. Their digital twins then combine first principle models and AI tools to host an idealised description of a physical manufacturing system represented by calibrated first principle, or hybrid models, and a digital replica of the actual plant.

Dr Lianlian Jiang, Co-Lead Principal Investigator of the project, Unit Lead of Digital & Sustainable Manufacturing at A*STAR I2R, said: “The AI agent in this ontology-based digital twin can be extended beyond anomaly detection to support quality monitoring, production scheduling and resource planning.

“By embedding domain knowledge into the system, the technology helps capture and transfer critical expertise while complementing staff expertise.”

Key advantages of an ontology-based digital twin, the partners said, are that it improves system understanding, refines predictions and supports the addition of flexible, new workflow agents for future digital development and manufacturing.

The technology will now be turned into a commercial product by another CARES spin-off, Chemical Data Intelligence, which will to make it available to pharma companies through the Pharma Innovation Programme Singapore (PIPS) Consortium.

In recent years digital twins have facilitated advanced process control productivity gains and been integrated with process analytical technology (PAT) in biologics manufacturing as they create environments where compliance information can be stored, digitised, and validated.