 |
 |
07 Sep 2023
Hewlett Packard Enterprise and the University of Rochester join forces under latest DOE scheme.
Researchers in the US are gearing up to employ machine learning and artificial intelligence (AI) techniques in a bid to improve the energy generated in laser-based inertial confinement fusion (ICF).
Armed with around $3 million in funding from the US Department of Energy (DOE), the three-year project sees scientists from the University of Rochester join forces with Hewlett Packard Enterprise (HPE).
They will use the OMEGA experimental database, generated by work at Rochester’s Laboratory for Laser Energetics (LLE), as training data for AI models, as well as simulation databases.
The plan is to then deploy the models to better understand the complexity of the underlying nonlinear physics of fusion, and ultimately to deliver a higher energy output from the imploding targets.
Real-time adjustments
In an article published on the Rochester web site, LLE chief scientist Riccardo Betti said: “Despite many years of laser-driven inertial confinement fusion research, there is not a clear path to the high-energy gains required for inertial fusion energy.
“However, we now have a wealth of experimental data that we can harness with machine learning to systematically correct the simulations and guide real-time adjustments to experiments.”
The DOE funding is part of a wider set of awards totaling $29 million aimed at improving fusion outcomes with AI and machine learning, but of the seven supported projects it is the only one focused directly on laser ICF.
Betti is listed as a principal investigator for the effort, entitled “Applications of Machine Learning and Data Science to predict, design and improve laser-fusion implosions for inertial fusion energy”, alongside Soumyendu Sarkar, a senior director and senior distinguished technologist in AI at HP Labs.
Also involved from Rochester are computer science associate professor Christopher Kanan, and LLE scientist Varchas Gopalaswamy.
Jean Paul Allain, DOE’s associate director of science for fusion energy sciences, said in a release announcing the seven projects:
“Artificial intelligence and scientific machine learning are transforming the way fusion and plasma research is conducted. These awards will advance a broad set of capabilities across the Fusion Energy Sciences (FES) program, making essential capabilities available for all stakeholders.
“The US is leveraging every tool in its pursuit of an aggressive program that will bring fusion energy to the grid on the most rapid timescale.”
 |  |  |  |  |  |  |
| © 2024 SPIE Europe |
|
