[object Object]

by The digiLab Team

Updated 7 March 2024

digiLab Joins U.S. DIII-D National Fusion Facility

Pursuing Advanced Plasma Control using AI
[object Object]

digiLab has begun efforts to advance fusion in the US, by joining the DIII-D National Fusion Facility, a U.S Department of Energy user facility. As a member of this facility that houses a world-class research tokamak, digiLab can now work with data and scientists from the facility to solve some of the hardest challenges in fusion. What's more, the experimental DIII-D facility can be used as a testbed for digiLab to explore its capabilities in machine learning for fusion.

Fusion at digiLab

Since its founding, digiLab has been engaged in difficult science and engineering problems that have positive impacts for the planet. And of all the technical challenges in the clean energy sector, perhaps the most difficult is nuclear fusion. Getting an economic fusion powerplant working necessitates solutions to extreme challenges in science and engineering; from achieving better core confinement through plasma physics, to optimising tritium breeding through materials engineering.

Figure 1. Predictions from a surrogate model for plasma turbulence developed by the UKAEA and digiLab. Surrogate models like these could be a powerful way to predict the performance of a fusion plasma, given properties like the plasma shape.

Figure 1. Predictions from a surrogate model for plasma turbulence developed by the UKAEA and digiLab. Surrogate models like these could be a powerful way to predict the performance of a fusion plasma, given properties like the plasma shape.

Collaborating with international fusion organisations, digiLab has already left its mark on the field of fusion. Leveraging its probabilistic machine learning and uncertainty quantification capabilities, along with an in-house team of fusion experts, digiLab has engaged in many areas of fusion development. One such area is in gyrokinetic models of turbulence (Figure 1), which helps optimise the performance of a fusion plasma. But digiLab has also worked on propagating uncertainties through complex engineering systems, to de-risk mission critical engineering components.

One particular focus of digiLab has been on improving the experimental control of fusion plasmas. Partnering with the UK Atomic Energy Authority, digiLab worked to automate the verification of gas controller inputs on the MAST-U tokamak.

The DIII-D Tokamak

Expanding on their current fusion collaborations, digiLab are now engaging the U.S. fusion community through the cutting-edge DIII-D facility, shown in Figure 2. DIII-D (short for Doublet III-D) is a leading tokamak experimental facility located in San Diego, California, and operated by General Atomics on behalf of the U.S. Department of Energy.

The DIII-D tokamak is undoubtedly a leading device in fusion, as it is the only tokamak of its class currently operating in the US. This also makes it one of the largest tokamaks currently in operation globally, and is one of the most well-diagnosed fusion machines, with of over 100 diagnostics. The DIII-D National Fusion Facility is also an incredibly productive centre, and it recently completed a single campaign that took place over 200 days, with more than 1,600 hours of plasma experiments.

Figure 2. A view inside the DIII-D tokamak (courtesy of General Atomics).

Figure 2. A view inside the DIII-D tokamak (courtesy of General Atomics).

Over nearly 40 years of history, the DIII-D facility has been a leader in many areas of experimental fusion physics. The facility pioneered work in advanced core confinement scenarios, which could allow for more stable, high-performance fusion plasmas. Moreover, the facility also leads research into plasma facing components, with projects testing materials to ensure they can survive the fusion environment without damage. Finally, in line with digiLab's previous work, the DIII-D programme is a pioneer in plasma control, having developed many tools to predict and prevent harmful events that could damage future machines. At the core all this work at the DIII-D National Fusion Facility is advancement of the science needed to deploy working fusion powerplants.

A Collaborative Space

Part of the success of the DIII-D programme is its ability to foster collaborative research and strategic partnerships in the field. DIII-D is home to over 700 users that can leverage DIII-D to publish high quality, open scientific works. The programme members include academic institutions; U.S. national labs such as the Lawrence Livermore National Laboratory and Oak Ridge National Laboratory; and private companies such as Next Step Fusion, Nvidia, and now, digiLab.

As a new member of the DIII-D programme, digiLab now has the opportunity to interact with the extensive datasets generated from tokamak experiments, and take part in the development of new experiments. Creating explainable, probabilistic machine learning methods for control will be a key focus of digiLab in the DIII-D programme. Specifically, digiLab is looking to develop tools to assess the uncertainties and risks in experimental control.

Ultimately, if these experimental risks can be identified, then work can be done to reduce and mitigate them. So whether digiLab is looking at the uncertainty in control actuator inputs, or the uncertainty in a plasma state due to diagnostics, the overarching goal is to become more confident in plasma control systems.

After all, tools to assess and optimise these uncertainties in control will allow fusion experiments to confidently push into the frontiers of science and engineering.

Interested in partnering or collaborating with digiLab to accelerate the commercialisation of fusion energy? Get in touch at fusion@digilab.co.uk

The digiLab Team
We are a 30+ strong team of ML Experts, Software Developers, Solution Engineers, and Product Experts. As a spinout from the University of Exeter, we build on years of cutting-edge academic research. At our core is a commitment to helping engineering and infrastructure companies become data-driven.

Featured Posts

If you found this post helpful, you might enjoy some of these other news updates.