Argonne National Laboratory has developed two artificial intelligence tools that tap into 60 years of nuclear fuel research to help design safer, more efficient reactor fuels. Rather than starting from scratch, these AI systems learn from historical experiments conducted at some of the nation's most important nuclear facilities, compressing decades of trial-and-error into predictive models that can guide the next generation of reactor development.

What Are These New AI Tools and How Do They Work?

Argonne's research team created two distinct AI systems, each designed to solve different problems in nuclear fuel engineering. The first, called "Dr. Metal," functions as a generative AI chatbot that acts as a virtual assistant with deep knowledge of metallic fuel research and historical datasets. Think of it as having a nuclear fuel expert available 24/7 who has memorized every relevant experiment from the past six decades.

The second tool, the "FGB Microstructure Generator," takes a more specialized approach. FGB stands for fission gas bubbles, which form inside fuel during reactor operation. This AI tool specifically predicts how these bubbles develop and change over time, a critical factor in understanding fuel performance under the extreme heat and radiation inside a reactor core.

Both tools leverage experimental data from major nuclear research facilities, including Chicago Pile-5, Experimental Breeder Reactor-II, and the Fast Flux Test Facility. By training on this historical data, the AI systems can identify patterns and relationships that might take human researchers years to discover through traditional analysis.

Why Does Predicting Fuel Behavior Matter for Nuclear Energy?

Understanding how nuclear fuel behaves under extreme conditions is fundamental to reactor safety and efficiency. Fuel must withstand intense heat, radiation, and pressure while maintaining its structural integrity. If engineers can predict fuel performance more accurately, they can design reactors that operate more safely and efficiently, which is especially important as the nuclear industry pivots toward smaller, modular reactors that require different fuel specifications than traditional large reactors.

The challenge has always been that obtaining accurate measurements and predictions requires extensive testing, which is time-consuming and expensive. By using AI to synthesize decades of existing research, Argonne is essentially compressing the knowledge acquisition phase, allowing engineers to move faster from concept to deployment.

How Are Researchers Improving Fuel Measurement Techniques?

Beyond AI tools, Argonne also recently validated a new measurement technique called the suspended bridge method, which addresses a long-standing challenge in nuclear fuel research. Thermal conductivity, or how well heat flows through fuel, is critical to understanding fuel behavior under reactor operating conditions. However, measuring thermal conductivity accurately during actual reactor operation has proven extremely difficult.

The suspended bridge method uses two tiny microfabricated platforms connected by an ultrathin sample to test thermal conductivity in a vacuum environment. This approach allows researchers to measure how various materials, including those used in nuclear fuel, conduct heat under controlled conditions that simulate reactor environments.

"This is a big step forward in understanding and optimizing nuclear fuel performance. It not only enhances reactor safety but also supports the design of next-generation nuclear systems," stated Abdellatif Yacout, a senior researcher at Argonne.

Abdellatif Yacout, Senior Researcher at Argonne National Laboratory

The suspended bridge method was developed through research supported by the Department of Energy's National Nuclear Security Administration, reflecting the government's commitment to advancing nuclear fuel science.

What Infrastructure Is Supporting This Research Acceleration?

Argonne also opened a new Activated Materials Lab (AML) located at the laboratory's Advanced Photon Source (APS), a major research facility. While Argonne has been researching irradiated materials at the APS for approximately 30 years, the lab previously lacked a specialized facility dedicated to this work. The new AML directly supports research on materials used in reactors and is funded by the Department of Energy Office of Nuclear Energy's Nuclear Science User Facilities.

The AML improves several critical aspects of fuel and materials research:

• Sample Accessibility: The new facility provides easier access to irradiated samples, reducing the time researchers spend on logistics and more time on actual analysis.
• Operational Flexibility: The lab's design allows for more flexible experimental protocols, enabling researchers to test a wider variety of conditions and materials.
• Cycle Time Reduction: By minimizing the time between sample preparation and testing, the AML accelerates the overall research timeline from concept to results.
• In Situ Testing Capabilities: The facility enables expanded real-time testing of materials, allowing researchers to observe how materials behave as they're being tested rather than only examining them afterward.

How Can These Advances Shape the Future of Nuclear Energy?

The combination of AI tools, improved measurement techniques, and dedicated research infrastructure represents a significant acceleration in nuclear fuel science. For the small modular reactor industry, which is gaining momentum as a potential solution to decarbonization, these advances are particularly valuable. SMRs require different fuel specifications and performance characteristics than traditional large reactors, and the ability to rapidly predict and validate fuel behavior could shorten development timelines considerably.

The AI tools also democratize expertise in a field where knowledge has historically been concentrated among a small number of experienced researchers. By encoding 60 years of research into accessible AI systems, Argonne is making it easier for the next generation of nuclear engineers to benefit from institutional knowledge that might otherwise be lost as experienced researchers retire.

As nuclear energy regains attention as a critical component of global decarbonization efforts, these research advances demonstrate that the nuclear industry is not simply relying on older reactor designs. Instead, it is actively modernizing its scientific foundation, using cutting-edge AI and measurement techniques to design safer, more efficient reactors for the 21st century.