The new AI tool can perform complex fusion plasma modelling in seconds instead of hours or days.

Computer scientists and fusion experts have developed the AI model that can create complex five-dimensional (5D) plasma turbulence simulations quickly and at a fraction of the cost. This, according to the UK Atomic Energy Authority, could pave the way for faster and more accurate design of future fusion power plants.

The tool, named GyroSwin, can create simulations up to 1,000 times faster than traditional computational methods.

“Designing, developing and operating a fusion power plant will involve millions of plasma simulations. Reducing runtimes from hours or days to minutes or seconds – whilst preserving sufficient accuracy – will be essential for making this challenge manageable,” said Rob Akers, director of computing programmes at UKAEA.

“Pioneering AI-based tools like GyroSwin therefore show great promise for being genuinely transformative around time-to-solution and cost.”

The UK Atomic Energy Authority said harnessing AI and supercomputing to model plasma turbulence faster – and at a much lower price – is essential towards fusion energy on the grid.

Fusion could be a cleaner and more abundant source of energy, but machines need to confine plasma at extreme temperatures using powerful magnets to achieve it.

A 5D approach requires immense supercomputing power, as traditional simulations are extremely slow and computationally expensive, significantly lengthening design and development cycles.

GyroSwin uses the latest AI methods to learn the 5D simulation dynamics and the resulting surrogate models can run in seconds. Such high levels of speed can enable much faster and more agile prediction of plasma turbulence, which is crucial for optimising fusion machine designs.

Processing 5D data has never before been tackled by an AI surrogate model and, according to the UK Atomic Energy Authority, GyroSwin outperforms other AI methods it has been compared against.

Johannes Brandstetter, Professor at JKU, co-founder and chief scientist at Emmi said: “We love scientific challenges and building AI models that accelerate 5D gyrokinetic simulations is definitely one of the toughest challenges out there.

“We are very proud of how far we got in this great collaboration, but we know that we have just scratched the surface.”

The UK Atomic Energy Authority is now expected to research how GyroSwin’s advanced capability can be applied to next generation power plants. As more complex physics is included for power plant conditions, simulations become even more lengthy, which makes faster plasma modelling essential.

Nuclear fusion has been heavily focused on in the data centre industry in recent months as a potential source of clean energy. Companies like Google and Amazon have already been looking into the use of nuclear power and other clean energies to power their data centres to confront rising emissions.

Related stories

Rolls-Royce wins UK’s first small nuclear reactor contract

ARC Clean Technology and Deep Atomic ink deal to power AI data centres with nuclear energy

Meta signs 20-year nuclear deal to power AI data centres from 2027