By The Financial District
Feb 241 min
Scientists dedicated to advancing fusion energy research claim they have made a significant breakthrough using artificial intelligence (AI), as reported by Angela Dewan for CNN.
Conducting their experiments at the DIII-D National Fusion Facility in San Diego, the team demonstrated that their AI controller could forecast potential plasma tearing up to 300 milliseconds in advance, thus averting a sudden halt to the fusion reaction. I Photo: Rswilcox Wikimedia Commons
For decades, nuclear fusion has been lauded as a nearly boundless source of clean energy, offering a potential game-changing solution to the climate crisis.
However, achieving and sustaining fusion energy has posed significant challenges, with experts only managing to maintain fusion reactions for a few seconds. Among the hurdles faced are instabilities inherent in the complex fusion process.
Various methods have been explored to achieve fusion energy, with one common approach involving the use of hydrogen variants as fuel input, along with the heating of plasma to extremely high temperatures within a donut-shaped apparatus called a tokamak.
Controlling this plasma is crucial, yet it remains highly vulnerable to "tearing" and escaping the powerful magnetic fields intended to contain it within the machine.
In a groundbreaking development, researchers from Princeton University and the Princeton Plasma Physics Laboratory disclosed in the journal Nature that they have employed AI to predict these potential instabilities and prevent them in real time.
Conducting their experiments at the DIII-D National Fusion Facility in San Diego, the team demonstrated that their AI controller could forecast potential plasma tearing up to 300 milliseconds in advance, thus averting a sudden halt to the fusion reaction.