Artificial intelligence brings us closer to realizing the promise of nuclear fusion

Achieving the promise of nuclear fusion as a source of limitless, safe, and clean energy is a grand challenge for science and engineering. The simplest closed magnetic geometry used to confine ultra-hot plasmas is the donut-shaped, rotationally symmetric tokamak. These configurations are susceptible to instabilities — tearing modes (TMs) — that reconfigure the magnetic field lines of the tokamak, break the tokamak’s symmetry, and can disrupt plasma operation. Much work is being devoted to resolving this.

Cristina Rea and Stuart Benjamin explored the expanding role of machine learning (ML) in helping achieve this goal.

“The mechanism by which TMs appear in tokamaks remains nonlinear, coupled, and chaotic,” said author Benjamin. “A balance of stabilizing and destabilizing effects at sensitive ‘rational surfaces’ where TMs form is tipped by intermittent, rapid instabilities elsewhere in the rotating plasma column. But the end state of an unmitigated tearing mode is simple — a great magnetic bubble grows within the plasma like a slug, grinding rotation to a halt, before dispersing the plasma into the wall.”

The researchers focused on various areas, including ML prediction of tearing mode onset, the use of ML to help interpret tearing onset data, and plasma controllers that leverage artificial intelligence-based tearing stability predictors.

“TMs remain fiendishly hard to predict with physics models, but their stochastic complexity appeals to ML-empowered scientists,” said Benjamin. “That’s why we wanted to explain how recent studies have applied AI to large experimental tokamak datasets, providing new insights into the TM physics and control mechanisms we must perfect to ensure TMs don’t compromise the tokamak power plants of the future.”

Source: “A review of machine learning-driven studies of tearing modes in tokamaks,” by C. Rea and S. Benjamin, Physics of Plasmas (2026). The article can be accessed at https://doi.org/10.1063/5.0325461 .

This paper is part of the 2nd European Conference on Magnetic Reconnection in Plasmas Collection, learn more here .