Dr Angel Jimenez-Aranda, Associate Professor of Digital Transformation at Salford Business School, recently presented his research on intelligent manufacturing systems at the 9th International Conference on Advanced Manufacturing and Materials (ICAMM 2025) in Oxford. The event brought together global experts to explore the latest advances shaping the future of manufacturing.
Addressing a critical industry challenge
As manufacturers continue to embrace digitalisation, they face ongoing challenges: unplanned downtime, equipment failure and overwhelming volumes of machine data. Traditional condition monitoring methods are often reactive, fragmented, and inaccessible, particularly for small and medium-sized enterprises (SMEs).
To address these challenges, a multidisciplinary research team at the University of Salford, led by Dr Tarek Gaber (School of Science, Engineering and Environment) and Dr Angel Jimenez-Aranda, developed SARAD (Secure AI for Robust Anomaly Detection). This innovative, explainable AI-based system helps manufacturers detect early signs of machine failure before disruption occurs, enabling them to act proactively rather than reactively.
What sets SARAD apart is its use of explainable AI. The system not only identifies unusual machine behaviour but also explains why the anomaly has occurred in clear, simple terms. If SARAD detects abnormal activity, it doesn’t just issue an alert, it offers understandable insights into what triggered the issue, supporting engineers, operators, and decision-makers in understanding and acting on the data.
Integrating AI with cybersecurity
To protect against emerging cybersecurity threats, SARAD includes secure AI countermeasures that are built directly into the system. These features enable the platform to detect and block attempts to manipulate input data, exploit vulnerabilities, or mislead the system into making incorrect decisions.
This security layer helps ensure that SARAD remains trustworthy and robust, even in high-risk digital environments. By integrating cybersecurity with AI, the system protects both the data and the insights derived from it, giving manufacturers peace of mind while using AI-based technology on the factory floor.
Practical, transparent and sustainable
Reflecting on the conference, Angel said: “ We wanted to create an innovative solution that works for the real world, especially for small and medium-sized manufacturers who don’t have the luxury of large data science teams. With SARAD, we’re showing how AI can be practical, accessible, transparent, and genuinely useful on the factory floor.”
SARAD allows manufacturers to reduce downtime, avoid unnecessary repairs, and optimise maintenance schedules. It also supports wider sustainability goals by reducing energy consumption, preventing material waste, and extending equipment lifespan.
Research aims
The SARAD project seeks to:
- Design and implement a semi-supervised machine learning solution to identify unusual or abnormal behaviour in manufacturing systems before faults occur.
- Enable predictive maintenance and support sustainability, helping manufactures to improve operational efficiency, reduce unplanned downtime, maintenance costs and energy consumption.
- Monitor equipment performance in real time, improving asset reliability and reducing wear and tear.
- Increase transparency and trust in AI systems by embedding explainability into anomaly detection processes.
- Ensure AI security and robustness by building in defences against adversarial attacks and data manipulation.
SARAD is a practical example of how academic research can have real-world impact. Its user-friendly design and powerful capabilities make it a valuable tool for manufacturers seeking to modernise their operations without overcomplicating their systems.
This research highlights the role digital transformation can play in driving productivity and resilience across the manufacturing sector, especially for SMEs looking to stay competitive in a fast-changing world.
For more details you can visit https://sustainable-innovation.salford.ac.uk/sarad/.