Nuclear engineers are getting excited about the cool things happening with artificial intelligence (AI) and machine learning (ML) lately. Even though the two have been getting better, there is a problem. We have not had specific tests to make sure they work well in nuclear engineering. So we cannot use them as much as we would like. Recognizing the need to establish a solid scientific and technical foundation for the development of next-generation nuclear systems, the Task Force on Artificial Intelligence and Machine Learning for Scientific Computing in Nuclear Engineering has been formed within the Expert Group on Reactor Systems Multi-Physics (EGMUP) of the Nuclear Science Committee’s Working Party on Scientific Issues and Uncertainty Analysis of Reactor Systems (WPRS).
The Task Force plans to create tests that focus on important AI and ML activities. These tests will cover different areas of computer stuff, from simple physics to lots of complex physics.
A significant milestone has been achieved with the successful launch of the first comprehensive benchmark for AI and ML in predicting Critical Heat Flux (CHF). CHF marks the limit in a boiling system beyond which wall heat transfer significantly decreases. It is often referred to as a critical boiling transition or boiling crisis. In a heat transfer-controlled system like a nuclear reactor core, it can lead to a substantial increase in wall temperature, potentially accelerating wall oxidation and causing fuel rod failure. Although it is crucial for reactor safety, accurately predicting it is challenging due to the complex dynamics of local fluid flow and heat exchange.
Current CHF models rely mainly on empirical correlations validated for specific application domains. The AI and ML methods utilized in this benchmark seek to enhance CHF modeling by directly leveraging a comprehensive experimental database provided by the US Nuclear Regulatory Commission (NRC).
The CHF benchmark’s phase 1 kick-off meeting on October 30, 2023, witnessed robust participation, with 78 attendees from 48 institutions across 16 countries. A lot of people from around the world getting involved shows that scientists everywhere are really interested and committed to using AI and machine learning in nuclear engineering.
