Artificial intelligence speeds up the creation of fusion energy

This video is a part of the FCAI success stories series. In the video series, we explain why fundamental research in AI is needed, and how research results create solutions to the needs of people, society and companies.

Fusion energy is no longer a project for the distant future. The current official EU aim is 35 years, and with AI methods, fusion is here even faster than what otherwise would be possible.

University of Helsinki has begun to lead a new center, E-TASC Helsinki Advanced Computing Hub, specialized in fusion research based on artificial intelligence and modelling.

“At the University of Helsinki and FCAI, we recently got major EU funding to take into use state-of-the art artificial intelligence tools to model fusion plasmas and materials. With these methods that we develop, one can speed up the development of energy-producing fusion power plants significantly”, says Professor of computational materials physics Kai Nordlund, dean of the Faculty of Science at the University of Helsinki.

The possibilities of fusion energy have been investigated since the 1950s. Unlike the nuclear reactors currently in use, which are based on fission, fusion reactors generate power by combining isotopes of hydrogen. Fusion would enable a form of energy production where the availability of the fuel is geopolitically neutral. Moreover, fusion energy does not generate any greenhouse gas emissions.

“With the methods we develop, one can speed up the development of energy-producing fusion power plants significantly”, says Kai Nordlund. Photo: University of Helsinki.

Fundamentally, fusion produces energy by the same mechanism as the sun: by fusing light atoms together to heavier ones, which releases millions of times more energy per atom than ordinary chemical processes.

“In the sun, this is made possible by a crushing pressure due to the enormous mass of the star and a heat of 10 million degrees. On earth, we cannot achieve similar pressures, so we have to use even higher temperatures, about 100 million degrees, and keep this minisun stable in a container”, says Nordlund.

This can already be done using electromagnetically controlled plasmas.

3D Render fusion reactor nuclear fusion, inside heated plasma, Tokamak.

“So far, these plasmas have been kept hot enough for energy production for only a few seconds. Achieving the long-time confinement needed is really difficult, and designing a reactor that can do this requires careful and reliable optimization of the reactor shape and control parameters”, explains Nordlund.

Combining AI expertise to other fields of science

The Finnish AI-support hub in fusion research is combining the strong AI expertise in the Helsinki region with other fields of science and to the Finnish industry.

The center brings together AI expertise from FCAI and the University of Helsinki, high-performance computing expertise from CSC, Aalto University’s expertise in fusion physics, the information technology expertise of Åbo Akademi University and VTT’s knowledge of fusion power plants.

Lumi, a supercomputer to be built in Kajaani, offers resources for performing the long and memory-intensive calculations required by fusion simulations and AI applications.

Thanks to the €3 million awarded for the period of 2021–2025, the new center will become the largest individual EUROfusion project ever established in Finland. The grant was awarded by EUROfusion, a consortium that coordinates fusion research in Europe.

The European Union aims to generate electricity from fusion energy by 2050 at the latest, with the EUROfusion consortium aiming to deploy a demonstration power plant already in 2040.

Further information:

Kai Nordlund
Professor of computational materials physics
Dean of the Faculty of Science
University of Helsinki
kai.nordlund@helsinki.fi
+358 29415 0007