Crossing the Digital Validation Chasm

A digital validation chasm exists between good manufacturing practices and the adoption of advanced control strategies, according to researchers at the Pharmaceutical Manufacturing Technology Center, University of Limerick. They propose bridging that gap with a unified, risk-based digital validation workflow that combines process understanding and model-based design with accepted quality risk management and lifecycle-oriented validation principles.

The chasm between traditional validation processes and digital technologies isn’t a regulatory issue, one of the researchers, Dermot McMorrow, PDEng, assistant professor, digital manufacturing, tells GEN. “Increasingly…it’s more about translating risk-based, lifecycle validation principles into day-to-day engineering practice,” he adds.

“Guidance supports the use of automation, iterative development, digital tools, and critical thinking…but validation is still often implemented through sequential, document-centric qualification activities that are fragmented from engineering development,” McMorrow elaborates. “The challenge, therefore, is organizational and methodological—how to integrate process understanding, risk assessment, engineering development, verification, and evidence generation into a coherent lifecycle.”

The model predictive controller the team developed for pH regulation used a data-driven Hammerstein-Wiener model to help tune the controller and “a process and control focused analysis…to identify and prioritize relevant failure modes and to inform requirements and verification scope,” according to the researchers, who were led by Damian McCarthy, PhD, staff robotics test engineer, and senior author Eoin Hinchy, PhD, associate professor, automated manufacturing.

“The main benefit,” McMorrow says, “is that the workflow connects process understanding, quality risk management, controller development, requirements, verification, and deployment within the same engineering lifecycle. Rather than treating validation-related verification as something performed after engineering is complete, evidence can be generated as the system is developed.

“The approach also creates explicit traceability from identified risks through functional requirements and acceptance criteria to verification activities and resulting evidence,” he adds.

Consequently, verification efforts can focus on the most relevant risks, thus supporting both automated repeatable testing and physical challenge testing. Because process changes can be validated relatively easily, one practical barrier to continuous system enhancements may be overcome.

The team also demonstrated in a real-world setting “that the controller developed and verified within the model-based environment can be deployed within an industry programmable logic controller environment,” McMorrow says. “Using the standard IEC 61131-3 structured text format also provides an industrially recognized representation of the controller implementation and supports traceability within the wider model-based workflow.”

Iterative imperative

Implementing such a workflow change, particularly in a highly regulated environment, should be done iteratively, with ample testing and user feedback.

Key implementation points, the researchers note, include assessing the maturity of the innovation to be implemented and the organization’s level of readiness for that advancement. The team’s prior work provides structured tools, such as an innovation maturity rubric, to guide those assessments.

They found that model-based design-driven digital validation processes mitigated digital validation challenges. Specifically, they “streamlined compliance tasks, reduced resource burdens, and enhanced traceability, demonstrating the framework’s practicality in real-world settings,” according to McMorrow, in his earlier doctoral research.

The team’s proposed workflow “has the potential to reduce reliance on repeated manual testing and documentation,” McMorrow says, and thus support the real-time reconfigurations needed in modern production environments.

But, McMorrow cautions, “Any industrial implementation would still need to be incorporated into the organization’s own quality system, validation strategy, and regulatory context. The study…is a practical proof-of-concept…rather than a pre-validation solution ready for direct deployment.” Related work by this team has also applied model-based validation approaches to new and existing equipment, including legacy programmable logic controller systems.

The post Crossing the Digital Validation Chasm appeared first on GEN - Genetic Engineering and Biotechnology News.