PPPL and Princeton develop AI framework to control fusion plasma in milliseconds

September 6, 2026 at 12:33 PM2 min

This digest was compiled by AI from multiple sources — links to the originals are below.

Researchers at the U.S. Department of Energy's Princeton Plasma Physics Laboratory and Princeton University have developed PACMAN, an AI software framework that controls fusion plasma in milliseconds. The framework was successfully tested on a real fusion system in five separate experiments. The design and initial results are described in a paper published in the journal Nuclear Fusion.

Key Facts

PACMAN stands for Prediction And Control using MAchiNe learning.
The framework was tested on a real fusion system in five separate experiments.
The design and initial results are published in the journal Nuclear Fusion.
Plasma instabilities can grow within milliseconds and disrupt the fusion reaction.
Advanced computer simulations can take days or even months to complete.

The Millisecond Challenge

In fusion systems, particles hotter than the core of the sun can become unstable within a few thousandths of a second. That is far too fast for a human operator to respond. For fusion to continue successfully, the plasma must remain hot, dense, and stable. Even relatively small disturbances, known as instabilities, can grow within milliseconds and disrupt the fusion reaction. Advanced computer simulations can take days or even months to complete, making them too slow for real-time control.

PACMAN Framework

PACMAN is a novel abbreviation for Prediction And Control using MAchiNe learning. The framework uses artificial intelligence to make rapid decisions while maintaining strict safety controls. People remain responsible for setting the system's objectives. PACMAN was designed to provide a shared structure that makes it easy for different machine learning models to work together. Fusion systems require multiple models because different parts of the machine and plasma must be monitored and controlled at the same time.

Machine Learning Speed

Co-lead author Hiro Farre Kaga, a graduate student in the Princeton Program in Plasma Physics, said machine learning models can describe plasma behavior very well. He added that machine learning models are the only way to model the plasma in millisecond times. The speed of these models is what's key for control. Machine learning has already shown considerable potential for controlling fusion plasmas. Many previous efforts were developed individually, without a common framework.