"PhD Studentship: Machine Learning approaches to improve the efficiency of fluid dynamics simulations"

University of Birmingham - Department of Chemical Engineering

This fully funded PhD opportunity sits at the cutting edge of computational modelling, artificial intelligence, and national-priority Defence research. Hosted at the University of Birmingham and funded through a UK Defence programme, the project tackles one of the most important challenges in modern engineering: how to dramatically accelerate high-fidelity computational fluid dynamics (CFD) simulations using machine learning.

CFD is a cornerstone of engineering across energy, aerospace, automotive, chemicals, and Defence, but its computational cost can be prohibitive. Many realistic simulations take weeks or even months to run, making detailed sensitivity analysis, uncertainty quantification, and rapid design exploration effectively impossible. This project aims to change that. By integrating state-of-the-art AI and machine-learning techniques with established CFD solvers, the student will help develop new approaches that markedly reduce simulation times while retaining physical accuracy and trustworthiness.

The initial application focuses on complex, high-speed gas-flow problems, but the tools and methods developed will be broadly transferable across sectors and disciplines. The core ambition is to create AI-accelerated simulation pipelines that allow engineers and scientists to explore design space, risk, and uncertainty in ways that are simply not feasible today.

A defining feature of this PhD is the level of support and training provided. You will be part of a large, supportive, and highly interdisciplinary research team spanning engineering, applied mathematics, computer science, and data science. While prior expertise in the areas of AI and/or CFD is beneficial, it is not expected. Instead, the project is designed to actively support the student in developing powerful, in-demand skills in CFD, numerical modelling, machine learning, and scientific programming. You will gain hands-on experience with industry-standard tools (such as OpenFOAM), alongside unique modelling and AI frameworks developed at the University of Birmingham and used in high-impact academic and industrial research.

The training provided will provide a valuable foundation for your future career - advanced modelling and AI skills are now foundational across engineering, technology, and data-driven industries. Graduates with deep expertise in simulation-accelerated AI are exceptionally well positioned for careers in Defence, aerospace, energy, advanced manufacturing, software, finance, and beyond, whether in industry, national laboratories, start-ups, or academia.

If you are excited by combining physics, computation, and AI to solve real-world problems of national importance, and want to graduate with a skillset that will remain valuable for decades, this project offers a rare and powerful opportunity.

Funding notes:
The successful candidate will receive a full EPSRC stipend, plus an additional £5k top-up from the industrial sponsor, equating to a tax-free annual income of £25k. All fees for the PhD will also be covered by the sponsor, with additional funding to support travel and other expenses.

For more information please contact Prof. Kit Windows-Yule