DeepMind Helps Physicists Control Plasma in Tokamak

the nuclear fusion Humanity’s energy needs could one day be met, by providing large amounts of clean electricity. However, the development of this technology still faces many obstacles. From atoms Fromhydrogen They are heated to very high temperatures to form plasma, which must release a lot ofenergy.

One problem is figuring out how to keep this plasma at a temperature above the surface the sun Long enough to extract electricity from it. This is where it comes in deep mindGoogle’s sister company specializes in artificial intelligence. In an article published in the magazine temper natureThe company has partnered with the Swiss Plasma Center to create artificial intelligence (AI) specializing in nuclear fusion.

Artificial intelligence to test different forms of plasma

To contain the interaction, scientists use generator coils from magnetic fields. It should be perfectly calibrated to obtain the desired shape and to ensure that the plasma does not touch the walls of the reactor. To test other configurations to improve the reaction, it is necessary to recalibrate everything, which is a very complex and time-consuming task.

DeepMind AI It was designed to recalibrate 19 magnetic coils inside the Tokamak Variable Configuration (TCV) in Switzerland. It is a deep reinforcement learning system based on a neural network. The AI ​​was first trained in a simulation by observing how each of the coils affects the shape of the plasma. She then had to try to calibrate the files to recreate specific shapes.

Two stable plasmas for 200 ms

The AI ​​was able to correctly reproduce several shapes including the D-shaped cross-section, similar to the one that will be used in the International Thermonuclear Experimental Reactor (eter) in a Construction in the Bouches-du-Rhône section which is planned to start operating in 2025. The scientists then validated the simulation results by testing the formations directly in the TCV.

The artificial intelligence allowed the researchers to test new forms of plasma, including “drops,” in which two distinct plasmas are present inside the tokamak simultaneously. The system managed to install them during the 200ms test. the nuclear fusion It presents a huge challenge to AI, which is often used for well-defined tasks, such as image processingor role-based admiration for games. Here, the system had to manage a continuous flow but it could not be measured continuously.

The results of this collaboration could influence the design of future tokamaks, or even accelerate the development of a viable fusion reactor. It remains to be seen if the AI ​​will be able to achieve similar results on a larger tokamak.