AI can prevent disruptions in fusion devices

A team of international scientists led by a graduate student Yichen Fu of the U.S. Department of Energy’s (DOE) Princeton Plasma Physics Laboratory (PPPL) has showcased how Artificial Intelligence (AI) used in self-driving cars can predict and avoid disruptions in fuels fusion facilities called tokamaks. This breakthrough can help in reducing and avoiding disruptions and thereby avoid damage to the tokamaks. Disruptions are the sudden bursts of energy stored in plasma that fuels fusion reactions.

“It’s fascinating to see that a relatively simple machine learning model could accurately predict the complicated behaviour of fusion plasma,” said Yichen Fu while speaking with the American Institute of Physics publication called SciLight.

Tokamaks run an increased risk of disruptions as researchers push against the operational limits of the facilities in order to maximise fusion power to recreate the powers of the sun and the stars on Earth. The new AI system allows the scientists to keep a close look on the limits while boosting fusion power. This technology is crucial for ITER, an international tokamak that is being built in France to demonstrate the practicality of fusion energy. Fusion reactions combine light elements in the form of plasma to make energy. Plasma is the hot charged state of matter composed of free electrons and atomic nuclei that makes up 99 per cent of the visible universe. If scientists can successfully create fusion, the world can have a virtually inexhaustible supply of clean energy to generate electricity.