The quest for nuclear fusion energy is increasingly being aided by advancements in machine learning and artificial intelligence, experts say.
Algorithms are being deployed to predict disruptions that could potentially occur during experiments and may help turn the long-discussed energy source—which is based on the same process that naturally powers the Sun and stars—into reality.
“In the last four or five years… you’re suddenly seeing that old idea—that fusion is great but infinitely far away—has gone away,” Daniel Kammen, professor of energy at University of California Berkeley, said in a workshop this week, Forbes reported.
Prof. Kammen credited modern computing for turning the production of fusion energy from “someone else’s lifetime to a next-generation project.” He said: “There are now people who are projecting small-trial fusion plants that couldn’t have been done before without higher computing.” Academic analysis suggests he is correct.
Teased as a new source of clean and virtually limitless energy, fusion is the reaction in which two atoms of hydrogen combine together, to produce energy. Recreating the process on Earth has proven complex and has been ongoing for decades.
Atoms of hydrogen have to be heated to extremely high temperatures so they form a plasma and “fuse” together. A large donut shaped device known as a tokamak uses magnetic fields to control the plasma long enough for fusion to occur.
These magnetic devices can become physically unstable during experimentation, but experts say AI-based algorithms can crunch vast amounts of data to identify risks.
In April 2019, research that emerged from the U.S. Department of Energy’s Princeton Plasma Physics Laboratory (PPPL), Princeton University and a Harvard graduate student applied deep learning tools to forecast any disruptions that could stall reactors.
The deep learning code was called the Fusion Recurrent Neural Network (FRNN). The findings of the unique study were published in the journal Nature.
“When plasma in a fusion experiment becomes unstable, it can escape confinement and touch the wall of the machine, causing severe damage and sometimes even melting or vaporizing components,” said researcher Julian Kates-Harbeck.
“If you could predict those escapes, or ‘disruptions,’ you could mitigate their effects and build in safety protocols that would cool the plasma down gently and keep it from damaging the machine,” said Kates-Harbeck, a physics Ph.D. student.
Researchers said their tool demonstrated the ability to predict true disruptions within the 30-millisecond time frame required by ITER, a global energy project that is aiming to complete initial plasma experiments using its tokamak in December 2025.
“We don’t have good strategies for completely avoiding these disruption events yet,” Kates-Harbeck said at the time. “The best thing we can do is to predict when they are going to happen so we can avoid most of their adverse effects.
“That might be, for example, by injecting neutral gas that cools the plasma before it smashes into the wall. But you can’t mitigate anything if you don’t know it’s coming.”
Experts from the Massachusetts Institute of Technology (MIT) said in 2018 physicists were aiming to create a device to safely produce fusion energy within a decade.
Last January, MIT noted another project was focused on designing a “prototype fusion device” to be operational within next 15 years—aiming to combat climate change.