Scientists at Princeton University are turning to artificial intelligence (AI) to fill in the blanks on plasma and, consequently, utilize it to unleash nuclear fusion energy one day.

They unveiled their AI-powered Diag2Diag tech in a paper published in Nature Communications and contextualized their breakthrough in a news release.

Lead author Azarakhsh Jalalvand likened AI's role in monitoring and controlling plasma to providing audio to a silent film. A tech can use the visual information on hand to accurately fill in the "audio" for the film. 

With respect to fusion, Diag2Diag works similarly by analyzing real-world data to generate accurate synthetic data that offers critical practical advantages for commercial use. 

"Fusion devices today are all experimental laboratory machines, so if something happens to a sensor, the worst thing that can happen is that we lose time," Jalalvand explained. 

"But if we are thinking about fusion as a source of energy, it needs to work 24/7, without interruption."

There is a reason why the rapidly developing field of nuclear fusion is the focus of so much scientific time and investment. 

If properly tapped into, fusion could generate nearly unlimited energy without creating long-lasting radioactive waste, as in nuclear fission, in a process that mimics how the sun generates heat.

To go from laboratory to widespread use, the Princeton team believes the experimental systems' diagnostics need to be leaner and more efficient. For instance, tokamaks, doughnut-shaped fusion systems, rely on diagnostics called Thomson scattering. 

The diagnostics, in their current form, can't keep up with plasma to maintain stability and peak performance, researchers say.

Should we be pouring money into nuclear fusion technology? Yes — it'll pay off 👍 It's worth exploring 🔬 Not from our tax dollars 💰 No — it's a waste 👎 Click your choice to see results and speak your mind.

"Diag2Diag is kind of giving your diagnostics a boost without spending hardware money," said the study's principal investigator, Egemen Kolemen. 

These impacts are heightened because AI can provide insight into the pedestal — the edge of the plasma. This difficult-to-measure area is crucial to maximizing energy generation, and Thomson scattering is the only diagnostic that can do so. 

By enhancing Thomson scattering, Diag2Diag reduces the number of diagnostics needed. Looking ahead, that can make reactors smaller, simpler, and more efficient. 

That translates to lower maintenance costs, fewer parts unrelated to energy generation, more space, and reduced risk of system errors.

The team's work also provides insight into controlling edge localized modes (ELMs) that threaten the reactor's inner walls. With AI, scientists gained an understanding of how techniques for flattening plasma and protecting the reactor functioned.

"Due to the limitation of the Thomson diagnostic, we cannot normally observe this flattening," principal research scientist Qiming Hu observed. "Diag2Diag provided much more details on how this happens and how it evolves."

Next, the team behind Diag2Diag believes it can be applied to other fields, such as spacecraft and robotic surgery. In terms of fusion, it has the potential to continue enhancing diagnostics and bringing commercial applications closer to reality.

"Diag2Diag could be applied to other fusion diagnostics and is broadly applicable to other fields where diagnostic data is missing or limited," Kolemen concluded.

Join our free newsletter for weekly updates on the latest innovations improving our lives and shaping our future, and don't miss this cool list of easy ways to help yourself while helping the planet.