The research team, led by Professor Sun Youwen of the HEFEI Science Institute at the Chinese Academy of Sciences, has developed two innovative artificial intelligence (AI) systems to increase the safety and efficiency of fusion energy experiments.
Their findings have recently been published in Nuclear Fusion and Plasma Physics and Controlled Fusion.
Fusion energy has the promise to provide clean and virtually infinite power. However, in the case of future reactors, they should work reliably, avoid dangerous phenomena such as intense and intense events that could damage the reactor, and accurately control the confinement state of the plasma to maintain high performance.
To address these challenges, the team created two specialized AI systems.
The first AI system acts as a predictor of destruction. We use the decision tree model to detect early warning signs of confusion caused by so-called “locked modes.” This is a common plasma instability. Unlike typical black box AI, this model is interpretable. Not only does it say something is wrong, it also explains why it refers to the physical signal behind that prediction. The test correctly shows 94% of the time early warnings, with the alert moving 137ms ahead of the confusion.
The second AI tool monitors the state of the plasma. Instead of using individual models to identify operational modes (such as L-mode and H-mode) and detecting edge localized modes (ELMS), researchers have developed a multitasking learning model in which both jobs are performed at once. This approach improves accuracy and robustness. Results: 96.7% success rate for real-time recognition of plasma conditions.
These AI tools not only improve reactor safety, but also provide a deeper understanding of plasma behavior. The technology developed in this work lays the key foundation for the intelligent control systems needed for next-generation fusion reactors.
/Public release. This material of the Organization of Origin/Author is a point-in-time nature and may be edited for clarity, style and length. Mirage.news does not take any institutional position or aspect, and all views, positions and conclusions expressed here are the views of the authors alone.
