DeepMind uses AI to control plasma in a nuclear fusion reactor

Inside a tokamak, hydrogen gas is subjected to intense heat and pressure until it becomes a plasma hotter than the Sun’s core, creating the perfect environment for nuclear fusion. Powerful magnetic fields confine this plasma, preventing it from touching the walls and sustaining the reaction, but that’s no small feat: each coil must react to each variation in the plasma, producing the perfect field to contain it and keep the reactor running.

Now, DeepMind and EPFL’s Swiss Plasma Center (SPC) have developed a novel approach leveraging artificial intelligence (AI) to create a unified neural network to control all magnetic coils, taking a remarkable step forward. in the pursuit of unlimited energy production.

Their work has been published in the journal Nature.

Tokamaks are doughnut-shaped machines used in nuclear fusion research and are likely to be the design that potential future fusion reactors will use to generate power if it ever becomes economically viable.

As the plasma swirls around the torus shape with a current equivalent to lightning, energy is carried away from the plasma field by the neutrons, which have no electrical charge and are unaffected by magnetic fields. These neutrons collide with an outer “blanket” where they are absorbed, heating the material and generating steam to power the turbines.

To control the magnetic field, previous versions of EPFL’s Variable Configuration Tokamak (TCV) used 19 individual algorithms for each coil, probing the plasma thousands of times per second and reacting accordingly to confine it.

This new approach uses an AI designed by DeepMind that controls the variable voltage of each coil to maintain specific plasma configurations. Before they could use it on the real deal, the researchers used it on a simulation specifically designed to replicate the tokamak’s systems.

Once trained, the algorithm could produce specific plasma configurations using the exact parameters required, and could even create various plasma shapes and maintain two plasmas in a simulation.

Then it was tested on the real tokamak, and the results were extremely promising, mimicking the success it had before in the simulation.

Some of the shapes the AI ​​could create using plasma. Image credit: DeepMind and EPFL

Of course, this is only one key to the ultimate fusion energy puzzle, but it’s a huge step in the right direction.

“Once that’s done, it’s not the end of the story. Then you have to make it a powerhouse,” said Gianluca Sarri, a professor at Queen’s University Belfast, UK, New Scientist reports.

“And this AI is, in my opinion, the only way forward. There are so many variables, and a small change in one of them can lead to a big change in the end result. If you try to do it manually, it’s a very long process.