Ãå±±ÂÖ¼é

Aerospace systems generate vast amounts of maintenance and operational data, from sensor streams to technical logs, yet much of it remains unstructured, fragmented, and underused. Hidden within these records are insights that could help engineers detect faults earlier, track system degradation, and make better-informed maintenance decisions. But how can we turn this complex information into something reliable, explainable, and actionable at scale?

Digital twins offer a promising foundation, but to truly support engineering decisions, they need to go beyond simulation and begin to interpret and reason about the systems they represent. This PhD project will tackle that challenge by developing intelligent methods that combine AI techniques such as language models that interpret technical text and knowledge graphs that map engineering relationships, together with deep domain expertise. These methods open new possibilities for extracting and connecting knowledge at scale. The goal is to enhance digital twins with the capability to interpret complex engineering data and deliver insights that are robust, adaptable, and applicable across complex, high-value, safety-critical domains. This research will contribute to shaping the next generation of predictive and explainable digital twins.

The core challenge this PhD will tackle is how to help digital twins make sense of complex, messy maintenance data and turn it into clear, useful insights for engineers. The aim is to move beyond data collection and develop intelligent methods that enable digital environments to understand and reason about real-world problems in safety-critical settings. To explore this, you will work with emerging AI techniques such as large language models, which can interpret technical documentation, and knowledge graphs, which help structure and connect engineering knowledge. These techniques will be combined into intelligent methods that support the development of reliable, AI-enabled decision-support tools to enhance maintenance planning and operational decision-making in complex aerospace environments. 

The objectives of the PhD are:
• Extract structured engineering knowledge from unstructured maintenance data using LLMs, and represent it using ontologies and knowledge graphs 
• Develop a standards-aligned semantic framework to ensure interoperability, reusability, and scalability across systems and sectors
• Model system degradation over time by developing temporal knowledge graphs and reasoning techniques to support predictive maintenance and asset health monitoring
• Design feedback mechanisms that deliver interpretable insights (e.g. alerts, recommendations, confidence scores) to engineers and automated systems
• Validate the system’s resilience, scalability, and practical relevance using real-world and representative datasets, with evaluation of technical performance and potential for wider industry adoption

Ãå±±ÂÖ¼é is a specialist postgraduate institution with a strong track record of applied research and close industry collaboration. The successful candidate will be based within the Manufacturing, Materials and Design theme at the Centre for Digital and Design Engineering (CDDE), which offers access to advanced simulation, visualisation, and high-performance computing facilities. The Centre supports research in digital twins, knowledge-based systems, AI, and immersive technologies such as VR and AR. The candidate will work independently and collaboratively with experts in the field, contributing to a dynamic, research-led environment. 

This project is sponsored by Rolls-Royce, a global leader in aerospace and defence innovation. The sponsor brings deep domain expertise, access to real-world data, and a clear pathway for the industrial application of research outcomes, ensuring strong alignment with current and future industry needs.

A major challenge facing industry today is how to turn vast, unstructured maintenance data into timely, actionable insight that supports safer, more efficient engineering decisions. This project will address that need by embedding intelligent reasoning and feedback mechanisms into digital twin environments, enabling them to interpret complex maintenance data more effectively. Using AI techniques, such as large language models, knowledge graphs, and ontologies, the research will help engineers transform unstructured maintenance records into explainable insights. By capturing both technical and experience-based (tacit) knowledge, the system will enable earlier fault detection, better understanding of system degradation, and more informed maintenance planning. Designed for scalability and resilience, the approach will integrate with existing workflows across design, manufacturing, and service, contributing to next-generation, data-driven maintenance strategies with strong potential for wider application in high-value, safety-critical sectors.

This funded PhD provides a unique opportunity to carry out applied research with direct industrial relevance, in collaboration with Rolls-Royce. Key benefits include:
• A 3-month placement at Rolls-Royce, offering practical experience in digital twin development, AI, and data-driven engineering.
• Opportunities to attend and present at national and international conferences, with full funding support.
• Specialist training in AI, machine learning, and digital engineering.
• Collaboration with academic and industry experts for technical insight and mentoring.
• A supportive research environment focused on both professional and transferable skill development, preparing graduates for careers in aerospace, engineering, and digital innovation.

Throughout the PhD, the student will develop a broad set of skills, from technical areas like digital twin development, machine learning, and system optimisation, to transferable skills such as communication, project management, and presenting complex ideas clearly. The real-world, interdisciplinary nature of the project will prepare them for diverse career paths in both industry and academia, especially within the fast-growing field of digital transformation.