Machine learning–enhanced boundary layer modelling for industrial CFD

Applications are invited for a fully funded Industrial Doctoral Landscape Award in partnership with Siemens Digital Industry Software, focused on advancing the next generation of industrial computational fluid dynamics (CFD). The project investigates how machine learning (ML) can be used to enhance the modelling of boundary layers in industrial CFD simulations, where complex geometries and computational constraints limit near-wall resolution. This PhD offers you the opportunity to conduct cutting-edge research with direct industrial impact combining fundamental fluid mechanics with modern data-driven techniques.

You will join a supportive team of over 50 researchers, technicians and academics within the Mechanical and Aerospace Systems research group and will have the opportunity to apply your research during a placement within Siemens Digital Industry Software.

Project overview

The project focuses on developing and integrating ML techniques to enhance wall treatments for under-resolved boundary layers in aerodynamic simulations for industrial applications. In many industrial settings, complex geometries and restricted computational resources make it impractical to generate sufficiently refined near-wall meshes, limiting the accuracy of conventional boundary layer modelling approaches.

During the PhD, you will curate an archive of high-fidelity simulation data spanning a range of representative application areas, which will be used to train and assess boundary layer neural network models. You will develop and evaluate suitable ML architectures, analysing the trade-offs between different modelling strategies and levels of fidelity. By the end of the project, you will demonstrate the integration of ML-based boundary layer models within an open-source finite volume CFD code and quantify their performance relative to current pragmatic industrial approaches.

You will spend at least three months during the PhD based within Siemens Digital Industry Software receiving joint supervision and training from both academic and industrial researchers and gaining direct exposure to industrial CFD workflows and software development practices.

Candidate requirements

We are seeking an enthusiastic, self-motivated researcher with a rigorous approach to problem-solving. You should have, or be expected to gain, a high 2:1 (preferably 1st class) honours degree in mechanical or aerospace engineering, or a related discipline with substantial background in fluid mechanics.

Essential skills

• Strong knowledge of numerical methods and fluid mechanics
• Experience with scientific programming and data analysis e.g. Python, Julia, MATLAB, C/C++, or similar
• Ability to work independently and as part of a collaborative research team

Desirable skills / experience

• Experience of applying CFD to a complex problem
• Appreciation of meshing requirements for aerodynamic simulations
• Experience with machine learning or data-driven modelling techniques

Eligibility and funding

Please note, due to funding restrictions this studentship is only available to UK (home fees) citizens.

This studentship covers UK home tuition fees and provides a tax-free stipend of £24,000 per year for four years. 

How to apply

This studentship is open until filled, however expected project start date is 1 October 2026.

To apply, you should email the following documents to Hadrian Moran at Hadrian.moran@nottingham.ac.uk

• Curriculum Vitae (CV)
• Cover letter
• Academic transcripts

Informal enquiries may be emailed to Dr Stephen Ambrose at Stephen.Ambrose3@nottingham.ac.uk