The Opportunity
As the U.S. pursues a diverse energy portfolio that produces less carbon, the need for large-scale, cost-effective deployment of nuclear power is apparent.
Advanced reactor designs leverage recent progress in materials, computational methods, and real-time diagnostics to address the needs for improved safety and reduced design, analysis, and deployment costs.
Traditionally, nuclear experts have used theory and empirical observations to create predictive models of nuclear reactor performance – a process which can be time consuming in regulatory frameworks that require high confidence in design and safety characterization. Simulations results are compared with real-world observations, models are adjusted, and simulations run again. This process continues until models are able to predict real-world behavior with an acceptable level of accuracy.
Recent technological advances – such as in artificial intelligence (AI) and machine learning (ML) – coupled with the desire to deploy reactors with longer operating lives and to increase the operational flexibility and safety of reactors, suggest that nuclear researchers should employ the latest technological tools to construct the best possible reactors.
What Argonne Offers
With over 50 years of investigating new designs for nuclear reactors and developing codes to analyze their safety and performance, researchers at the U.S. Department of Energy’s Argonne National Laboratory are now integrating that knowledge with AI and ML tools.
Researchers in Argonne’s Nuclear Science and Engineering division are using ML methods specifically to generate fast-running models and improve predictive capabilities. By using AI and ML, researchers can develop computational methods to create a framework that: supports rapid and comprehensive design; supports efficient analyses that probe the entire design and operation domain; and improves characterization of safety margins by reducing uncertainties – all of which contribute to a reduction in nuclear power cost from the beginning of the design process to long-term operation.
Argonne’s unparalleled nuclear history and long track record with industry – coupled with being leaders in AI and having high-performance computing capabilities – allows for further advancement of new nuclear reactors.
The Benefits
Utilizing AI and ML for advanced reactor design offers multiple advantages:
- An improved process for advanced nuclear reactor design and analysis
- Improved characterization of safety margins
- More accurate models in fewer iterations
- Efficient and faster analysis framework
- Reduced cost during the design process and operation
- Aid in re-establishing the U.S. as a leader in nuclear energy