Mechanical Engineering, Fusion, Digital: Digital twin validation - Physics-based model, AI and Data
About the Project
Key Information
Open to: UK and international applicants
Funding Providers: Foster, UKAEA and Faculty of Science and Engineering
Subject Area: Mechanical Engineering, Fusion, Digital
Project Start Dates: October 2026
Supervisors: Professor P. Nithiarasu, Dr Alberto Coccarelli, Lloyd Fletcher (UKAEA)
Aligned programme of study: Mechanical Engineering, Ph.D.
Mode of study: Full-time
Place of study: Swansea University (Bay Campus)
Project description
This project addresses a fundamental question in digital twin development for fusion energy systems: the relative roles and limitations of physics-based models, experimental data, and AI-driven approaches. The research will focus on fundamental thermomechanical and thermofluid phenomena relevant to fusion devices and components. In particular, it will investigate uncertainties arising from integrating experimental measurements into physics-based models, as well as the feasibility of replacing such models with AI systems trained either on experimental data or synthetic datasets.
Current literature is dominated by AI models trained on synthetic data, yet their accuracy and reliability under real operating conditions remain largely untested. Evaluating the fidelity of these models against experimentally informed physics-based approaches is therefore both necessary and timely.
By systematically quantifying uncertainties across physics-based and AI-driven frameworks, the project will provide a rigorous validation of existing methodologies used to construct digital twins. The outcomes will clarify when physics-informed modelling is essential, when data-driven approaches are sufficient, and how hybrid strategies can be used to deliver reliable digital twins for fusion energy applications.
Find Your Best Opportunity
Tell them AcademicJobs.com sent you!
