Net-zero aerospace targets [1,2] have been a dramatic engineering driver to 
increased combustion and turbine entry temperatures in gas turbine systems with 
kerosene and next-generation hydrogen combustors. In turbine and combustion 
systems within gas turbines, near-wall jet flows are present in a range of 
cooling applications. Accurately modelling near-wall jets flows in Computational 
Fluid Dynamics (CFD) is important in understanding the performance and lifetime 
of turbine and combustor components where Han et al. [3] report that a 2% error 
in predicted metal temperature can halve the component lifetime. However, 
current Reynolds-Averaged Navier-Stokes (RANS) turbulence models provide large 
uncertainties in the adiabatic wall temperature. 

Physics-Informed Machine Learning (PIML) [4] combines machine learning 
algorithms with the physical equations for the modelled phenomena. In RANS 
applications PIML can be used to train functional models of turbulent closures 
using training data from high-fidelity LES data and sparse experimental 
datasets. The incorporation of both high-fidelity computational data with 
experimental data is a novel area of interest and should be investigated and 
incorporated into the model. In addition, the PhD project should investigate 
interpretability methods of deep machine learning methods. The PhD aims to 
produce understandable machine-learnt turbulence models that provide improved 
predictive accuracy at low cost to a range of near-wall jet cases applicable to 
the gas-turbine industry. Furthermore, the PhD will provide a PIML strategy that 
can be extended to other engineering applications.

The results of this PhD will be disseminated at conferences and high-quality 
journals. Suitable journals include the Journal of Fluid Mechanics, Physics of 
Fluids, Journal of Computational Physics, Computers and Fluids and the 
International Journal of Heat and Mass Transfer. This project and its aims align 
with the research output and key areas of growth in the Mechanical and Aerospace 
Systems (MAS) research group while encouraging local collaboration between the 
Faculty of Engineering and Computer Science.

The selected PhD candidate will be expected to undertake training in the use of 
OpenFOAM, an open-source CFD suite, alongside C++ and Python languages. 
Additionally, machine learning training will be advised. These skills are both 
necessary for the project but also provide an important set of professional 
skills for their future career. 

[1] UK Department for Transport. Jet Zero strategy: our approach for achieving 
net zero aviation by 2050. [Government of the United Kingdom], (2022, July 19). 
Retrieved from 
https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attac
hment_data/file/1095952/jet-zero-strategy.pdf
[2] European Commission, Directorate-General for Mobility and Transport, 
Directorate-General for Research and Innovation, Flightpath 2050 : Europe’s 
vision for aviation : maintaining global leadership and serving society’s needs. 
Publications Office, (2011). Available from: doi/10.2777/50266.
[3] Han, J.C., Dutta, S. and Ekkad, S., Gas turbine heat transfer and cooling 
technology, CRC press, (2012)
[4] Karniadakis, G.E., Kevrekidis, I.G., Lu, L. et al. Physics-informed machine 
learning. Nat Rev Phys 3, 422–440 (2021).