As the urgency of addressing climate change intensifies, fusion energy emerges as a promising clean energy solution. Despite decades of laboratory research on fusion reactions, several crucial questions remain unanswered to make fusion power a reality. In pursuit of carbon neutrality by 2050, the U.S. Department of Energy (DoE) has allocated new funding for a project led by researchers at MIT’s Plasma Science and Fusion Center (PSFC) and four collaborative institutions.

Cristina Rea, a research scientist and group leader at PSFC, assumes the role of primary investigator for this three-year collaboration aimed at integrating fusion data into a format compatible with AI-powered tools. Working in conjunction with scientists from William & Mary, the University of Wisconsin at Madison, Auburn University, and the nonprofit HDF Group, the project’s objective is to create a comprehensive fusion data platform that fosters inclusivity in research and data science. The initiative prioritizes outreach programs led by co-investigators, with four out of five being women, to encourage diverse participation in fusion research.

The DoE’s funding, part of a $29 million investment in seven projects spanning 19 institutions, will support the distribution of data generated by fusion devices, including the PSFC’s Alcator C-Mod, a “tokamak” using powerful magnets to control and confine fusion reactions. The data from Alcator C-Mod, which operated from 1991 to 2016, continue to be studied, thanks to the PSFC’s commitment to open knowledge exchange.

Currently, data from nearly 50 public experimental magnetic confinement-type fusion devices can be challenging to access due to varying cataloging methods and data organization. This hampers the use of machine learning and AI tools for data analysis and scientific discovery. Consequently, fewer scientists engage in fusion research, leading to barriers in discovery and a slowdown in harnessing AI’s potential to accelerate progress.

The proposed data platform aims to overcome these technical barriers by adhering to FAIR principles (Findable, Interoperable, Accessible, Reusable) and UNESCO’s Open Science (OS) recommendations, ensuring transparency and inclusivity in science. All deliverables will adhere to FAIR and OS principles, as mandated by the DoE. The platform’s databases will be constructed using MDSplusML, an upgraded version of the MDSplus open-source software developed by PSFC researchers in the 1980s to catalog Alcator C-Mod’s experiment results. MDSplusML continues the tradition of open collaboration, with nearly 40 fusion research institutes using MDSplus to store and provide external access to their fusion data.

To enhance accessibility and participation for women and underrepresented groups, the project will organize a subsidized summer school focusing on fusion and machine learning, hosted at William & Mary for the next three years.

Cristina Rea underscores the significance of their research, stating, “This project responds to the fusion community’s needs and prepares us for success. Multidisciplinary collaboration and diversity are essential for scientific advancements in fusion, as they accelerate problem-solving.”

This collaborative effort aligns with key research areas identified in the International Atomic Energy Agency’s “AI for Fusion” Coordinated Research Project (CRP). Rea, selected as the technical coordinator for the IAEA’s CRP, emphasizes community engagement and knowledge access to expedite fusion research and development. The IAEA acknowledges the global benefits of the researchers’ work, stating it will benefit both their CRP and the international fusion community.

Dennis Whyte, PSFC Director and Hitachi America Professor of Engineering, expresses his enthusiasm for the project’s role in applying AI tools and facilitating critical data extraction from experiments.

Cristina Rea concludes, “Leading such an important project is a meaningful opportunity, and I feel a responsibility to demonstrate that women are leaders in STEM. Our motivated team is committed to enhancing the fusion ecosystem and contributing to the realization of fusion energy.”

By Impact Lab