Scientists at Princeton University have announced a historic breakthrough after solving a puzzle to overcome the huge challenge of producing near-limitless clean energy with nuclear fusion.
Previously, the scientific community believed conquering nuclear fusion was decades away.
However, scientists have now revealed that they have used artificial intelligence (AI) to crack the puzzle.
The team of American scientists figured out a way to use an AI model to predict and prevent instabilities with plasma during fusion reactions.
Nuclear fusion has been hailed as the “holy grail” of clean energy for its potential to produce vast amounts without requiring any fossil fuels or leaving behind any hazardous waste.
The process mimics the same natural reactions that occur within the Sun, however harnessing nuclear fusion energy has proved immensely difficult.
In 2022, a team from the Lawrence Livermore National Laboratory in California achieved the first-ever net energy gain with nuclear fusion, meaning they were able to produce more energy than was input for the reaction, as Slay News reported at the time.
It was only a small amount – roughly enough to operate a small household appliance for a minute or two – but it represented a major milestone towards achieving it at scale.
The newly developed AI model operates by learning from real fusion experiment data, allowing it to adjust power input and plasma shape to mitigate instabilities.
Tested successfully at the DIII-D National Fusion Facility in San Diego, this AI application not only represents a significant technical achievement but also illustrates the potential for AI to play a pivotal role in making fusion energy a practical reality.
By addressing the tearing mode instabilities, the researchers are paving the way for sustained high-power fusion reactions, a crucial milestone for efficient energy production.
While the research is still in its early stages, the implications of this advancement are far-reaching.
The ability to control plasma more effectively opens up new avenues for enhancing reactor safety and efficiency.
Moreover, this breakthrough underscores the broader potential for AI applications in fusion energy technology.
As the AI model continues to learn and improve, it could lead to more robust solutions for stabilizing plasma in tokamak reactors and beyond, marking a significant step forward in the quest to unlock the full potential of nuclear fusion energy.
The latest success means another significant obstacle has been passed, with the AI capable of recognizing plasma instabilities 300 milliseconds before they happen – enough time to make modifications to keep the plasma under control.
The new understanding could lead to grid-scale adoption of nuclear fusion energy, according to the researchers.
Research leader Egemen Kolemen, who works as a physicist at the Princeton Plasma Physics Laboratory where the breakthrough was made, said the technology can be upscaled to provide near-limitless energy around the world.
“By learning from past experiments, rather than incorporating information from physics-based models, the AI could develop a final control policy that supported a stable, high-powered plasma regime in real-time, at a real reactor,” said Kolemen.
The latest research was published in the peer-reviewed scientific journal Nature on Wednesday in a paper titled “Avoiding fusion plasma tearing instability with deep reinforcement learning.”
“Being able to predict instabilities ahead of time can make it easier to run these reactions than current approaches, which are more passive,” said SanKyeun Kim, who co-authored the study.
“We no longer have to wait for the instabilities to occur and then take quick corrective action before the plasma becomes disrupted.”
