Oxford researcher leads new thermofluids–AI programme for smarter AIDC cooling

Dr Muting Hao, Associate Member of Faculty in Engineering Science and Career Development Research Fellow at St John’s College, developed the research direction and is leading the resulting programme at the Oxford Thermofluids Institute - the only Engineering team in the current Oxford–Cambridge–Imperial QRT Labs cohort.

As AI use grows, data centres need more effective ways to keep their powerful computers cool. The amount of heat generated can change rapidly as computing demands fluctuate, making it difficult for conventional computer models to predict cooling needs quickly enough or support fast operational decisions. More responsive modelling and cooling strategies could therefore play an important role in reducing energy use while maintaining reliability and enabling higher-density AI computing.

To address this challenge, Dr Muting Hao, is combining engineering models with AI to predict how heat moves through complex systems and help engineers make faster decisions. DPhil student Mr Jierui Li has been awarded an Oxford × QRT Scholarship, to undertake a project within this programme (representing around £200,000 of support).

Supported by the Scholarship, DPhil student Jierui Li will investigate how AI can help dynamic thermofluid modelling in different thermofluid systems, for example how AI can help cooling systems respond to changing computing demands. The aim is to reduce wasted energy on cooling while keeping data-centre equipment at safe operating temperatures. The research could help strengthen the UK's AI infrastructure by advancing more energy-efficient and reliable liquid-cooling technologies for next-generation data centres.

Dr Hao’s research team at the Oxford Thermofluids Institute develops computational methods for complex, dynamic thermofluid systems. Building on computational fluid dynamics, conjugate heat transfer and reduced-order modelling, her research combines physical modelling with AI and machine learning to support rapid prediction, optimisation and engineering decision-making. Applications span AI data-centre cooling, dynamic energy systems and aerospace propulsion.

The programme forms part of the wider QRT Labs network spanning Oxford, Cambridge and Imperial College London, bringing together researchers working across AI, mathematics, computing and quantitative science. Dr Hao’s team represents the only Engineering team in the current cross-university cohort, with Jierui as its only Engineering DPhil student. This gives the work a distinctive role in bringing physical engineering problems into the wider mathematical and computational AI research community. Dr Hao said:

“I am delighted that QRT’s support recognises the potential of bringing AI and dynamic thermofluid research together, and to see this new research direction taking shape. Our aim is not simply to make engineering models faster, but to do so without losing the physical understanding that makes their predictions trustworthy. In the modern By combining machine learning with governing physics, we hope to develop computational tools that can ultimately support real-time engineering decisions.

I am particularly pleased that this programme brings thermofluid engineering into the wider QRT Labs community and creates new opportunities for interaction across engineering, mathematics and computer science.”

Jierui will undertake his DPhil under the supervision of Dr Muting Hao and Professor Luca di Mare, whose expertise includes computational fluid dynamics . The scholarship also connects the project with researchers across Oxford, Cambridge and Imperial, as well as researchers and technologists at QRT. Jierui said:

“I am delighted to have been awarded the Oxford × QRT Scholarship, and I am deeply grateful to QRT, the Department of Engineering Science and my supervisor, Dr Muting Hao, for their generous support in making this possible. I am excited to be starting my DPhil at Oxford and really looking forward to the years of research ahead.”