EUROfusion awards to FinnFusion two spearhead projects in Artificial Intelligence and Machine Learning

By launching 15 new research projects, EUROfusion is engaging data science experts across Europe to apply Artificial Intelligence and Machine Learning techniques to fusion energy. FinnFusion / VTT will deliver two of the projects. These projects will leverage the world’s largest and most diverse dataset of fusion experiments to identify optimal methods for understanding and controlling the fusion process, ultimately shortening the road to energy applications.

Fusion energy promises to deliver safe, sustainable, and low-carbon baseload power, complementing other clean energy sources like solar and wind. To achieve this, we need to address complex physics and engineering challenges, including understanding the collective movements of charged particles in magnetic fields, mitigating disruption events, analyzing material erosion effects, and processing data rapidly enough for use in control loops. Artificial Intelligence and Machine Learning offer new opportunities to deepen our understanding of these phenomena.

Building on the world’s largest dataset

“With new research projects on Artificial Intelligence and Machine Learning, EUROfusion aims to accelerate progress towards fusion energy and support the ongoing efforts in its work packages”, explains Sara Moradi of the EUROfusion Programme Management Unit. “Machine learning and Artificial Intelligence are powerful tools for extracting insight from data, uncovering patterns and suggest control schemes that are too computationally intensive to identify with traditional computer models.”

Artificial Intelligence for fusion projects

Following a call and selection process, the EUROfusion General Assembly approved the support for 15 new Artificial Intelligence Fusion research projects. The strong response to the call highlights the scientific community’s commitment to using state-of-the-art approaches to advance computational techniques for magnetically confined plasmas. The 15 projects will receive a total amount of €2.659 million, of which half is provided by collaborative co-funding from the researchers’ home institutes and half from EUROfusion. The research projects will run for a period of two years. EUROfusion introduced all supported projects in the EUROfusion news.

Machine learning accelerated pedestal MHD stability simulations

Principal investigator: Dr. Aaro Järvinen, VTT

The performance of tokamak magnetic confinement fusion devices is strongly dependent on the achievable fuel pressure at the edge of the plasma, the ‘pedestal’ region. The overall edge plasma pressure is governed by combination of plasma transport and overall stability of the edge pressure profile. Therefore, repeated stability calculations are a typical feature of numerical prediction workflows for the pedestal region. These kinds of calculations can already be done with physics codes, but they are computationally heavy. The goal of this project is to substantially accelerate these tools so they can be used in fast everyday data analysis as well as in real-time applications like control.

Using machine learning to get faster simulations of the stability of the edge plasma (‘pedestal’) in tokamaks. Credit: A. Järvinen et al., VTT.

Applying Artificial Intelligence/Machine Learning for Neutral Beam Injection ionization and slowing-down simulations using ASCOT/BBNBI

Principal investigator: Dr. Antti Snicker, VTT

One of the main heating systems for many magnetic confinement fusion devices is based on injection of energetic, neutral particles to the plasma (Neutral Beam Injection). As these particles are ionized and collide with the plasma particles, they deposit their energy, heating the plasma. However, it is important to know where these particles deposit their energy and what is the fraction of neutral particles travelling through the plasma and hitting the opposite wall. These calculations can be conducted with numerical tools, but the calculations are computationally heavy. In this project, the aim is to substantially accelerate these calculations for fast everyday data analysis and real-time applications.

Fast ions tend to form complicated patterns, including hot spots, when they hit the plasma-facing components after interacting with the magnetic fields and plasma. The picture illustrates wall loads that have been assessed computationally with the ASCOT code.

About FinnFusion

FinnFusion is an open community encompassing all Finnish players in fusion energy business, technology and science, including funding agencies and ministries.

FinnFusion contributes to the realisation of fusion energy. The road to fusion is full of industrial opportunities for FinnFusion members. Besides projects, the members contribute with their expertise by participating in the FinnFusion advisory board, which meets regularly twice a year.