Civil Engineering : Fully Funded EngD Engineering Studentship in A real-time PINN-based computational framework for data integration and digital twinning (RS828)

Funding providers: Faculty of Science and Engineering and Fusion Engineering CDT in association with UKAEA 

Subject areas: Civil Engineering, Mechanical Engineering and Computer Science 

Project Start Date: 1 October 2025 

Supervisors: Professor Perumal Nithiarasu (Swansea University), Dr P Sharma (UK Atomic Energy Authority) 

Programme of Study: EngD Engineering  

Mode of Study: Full time  

Place of Study: Swansea University - This EngD project is funded by the Fusion Engineering CDT and hence the student will be based at Swansea University but should expect to engage fully with the 3-month training programme within the Fusion Engineering CDT at the start of the course. CDT training will be delivered across the CDT partner universities at Sheffield, Manchester, Birmingham and Liverpool. For further information about the CDT programme.

Project Description: 
The engineering components of the fusion process have a significant number of unknowns. We anticipate that significant unknown issues such as material properties, thermal stress behaviour and thermal behaviour will be uncovered as we build fusion power plants. The data will come in sparse measurements of material properties, temperature, and strains. Integrating such properties into a modelling platform is challenging as constructing the problem from sparse measurements requires some form of rapid inverse work to determine, for example, the property distribution across the entire component. This will be vital at the development stage of essential components such as the breeder blanket.  Inverse modelling comes with uncertainty and multiple solutions. Our previous fundamental work demonstrates that PINNs can help here but require further research.  

Although the PINNs can be expensive in complex problems compared to traditional computational methods, they provide two distinctive advantages when dealing with the complexities of the fusion component analysis. PINNs can use a unified platform for both inverse and forward analysis, and parameterisation can be carried out in one calculation, i.e., additional parametric variations of material properties can be incorporated as an additional feature. These two advantages are worth investigating in the context of the complexities of a breeder blanket.  

The following objectives are the objectives for this EngD project:  

• Year 1: Training and deep dive into data integration, literature survey and analysis of current state of the art of PINNs for combined forwards and inverse framework.  
• Year 2: Implement a unified inverse and forward framework on fundamental problems and test the method for speed, accuracy and robustness. 
• Year 3: Implement the framework into various blanket design analyses (current student’s work) and incorporate any available data.  
• Year 4: Comprehensive testing of both the combined inverse and forward PINN framework on competitive blanket designs, introducing PINN-based parametric studies and thesis writing.  

This project is at the interface between UKAEA engineering and computing groups. Swansea will act as the bridge between these groups to deliver the best design options by combining high-fidelity simulations with design needs. 

For more details please see here: https://www.swansea.ac.uk/postgraduate/scholarships/research/fse-fusion-ukaea-phd-engd-rs828.php