New AI method speeds up calculations to protect fusion reactors from plasma heatGeorgina Jedikovska
Interesting EngineeringWed, August 13, 2025 at 2:29 PM EDT
3 min read
Scientists in the US have introduced a novel artificial intelligence (AI) approach that can protect fusion reactors from the extreme heat generated by plasma.
The new method, which is called HEAT-ML, was developed by researchers from Commonwealth Fusion Systems (CFS), the US Department of Energy's (DOE) Princeton Plasma Physics Laboratory (PPPL), and Oak Ridge National Laboratory.
It is reportedly capable of quickly identifying magnetic shadows, which are critical areas protected from the intense heat of the plasma, and therefore help prevent potential problems before they start.
Locating these regions quickly and accurately is crucial for ensuring the long-term operation of fusion systems, where plasma temperatures can soar higher than the core of the Sun.
The researchers believe that the new AI could lay the foundation for software that significantly speeds up the design of future fusion systems and supports informed decision-making during operations by adjusting the plasma.
An innovative methodFusion, which is the reaction that powers the sun and stars, could supply Earth with limitless, carbon-free energy. Still, in order to achieve this, researchers must first overcome significant scientific and engineering challenges.
One of the biggest is managing plasma heat, which exceeds the temperature of the sun's core when confined in a tokamak, a donut-shaped reactor that uses powerful magnetic fields to contain the plasma.
That's why speeding up the calculations that predict where this heat will strike, and which parts of the tokamak will remain in the protective shadows of other components, is essential to bringing fusion power to the grid.
"The plasma-facing components of the tokamak might come in contact with the plasma, which is very hot and can melt or damage these elements," Doménica Corona Rivera, an associate research physicist at PPPL and first author on the paper, said.
To address the challenge, the researchers developed HEAT-ML as an AI-powered upgrade to the open-source Heat flux Engineering Analysis Toolkit (HEAT). The software generates 'shadow masks'. These are 3D maps showing which parts of a tokamak's inner walls are shielded from direct plasma contact.
The AI was designed specifically for SPARC, the tokamak under construction by CFS in Massachusetts, which aims to demonstrate net energy gain by 2027. For its initial testing, researchers focused on 15 tiles near the bottom of SPARC's exhaust system, the area expected to endure the most extreme heat.
Simulating fusion system performanceTraditionally, HEAT traces magnetic field lines from a component's surface to determine whether they intersect other internal structures, and marks those regions as 'shadowed.' While accurate, the process is slow, with a single simulation taking up to half an hour. It can take even longer for complex geometries.
The new HEAT-ML overcomes this bottleneck with a deep neural network trained on roughly 1,000 SPARC simulations generated by HEAT. Once trained, the AI can produce shadow masks in just a few milliseconds, cutting computation times by several orders of magnitude.
"This research shows that you can take an existing code and create an AI surrogate that will speed up your ability to get useful answers, and it opens up interesting avenues in terms of control and scenario planning," Michael Churchill, head of digital engineering at PPPL and study co-author, said in a press release.
Although the current version is tailored to SPARC's exhaust system, the team aims to expand its capabilities to handle exhaust systems of any shape or size, as well as other plasma-facing components within a tokamak.
The study has been
published in the journal Fusion Engineering and Design.
https://www.yahoo.com/news/articles/ai-method-speeds-calculations-protect-182943045.html
