In our world of AI-dominated algorithms, some nuclear physicists prefer to do science the old-fashioned way. And this newly developed physics-based, AI-free algorithm may be the breakthrough needed for small nuclear reactors to take off in earnest.
A recent paper published in Progress in Nuclear Energy describes an algorithm that enables nuclear microreactors—transportable, mini reactors supplying energy to remote locations—to automatically adjust their power output. This is particularly relevant to microreactors, as their intended placement in rural communities, disaster zones, and cargo ships means there won’t always be staff available for manual checkups, unlike conventional reactors.
“Our method can help vendors design reactors with autonomous control systems that are safer and more secure,” Brendan Kochunas, study senior author and a nuclear engineer at the University of Michigan, said in a statement.
“Many startup and legacy companies in the U.S. are pushing towards near-term and broad deployment of nuclear microreactors,” he added. “And our work establishes a clear avenue to achieve that in an economically viable way.”
Small but powerful
Indeed, microreactors are a powerful yet mobile option for supplying electricity to remote areas. To be clear, microreactors are different from small modular reactors, which are scaled-down versions of conventional nuclear reactors. Microreactors have a smaller power capacity that generates up to 20 megawatts of thermal energy.
In the United States, microreactors have been around as early as the mid-20th century, mostly for military use. With advances in nuclear physics, microreactors began to appear in more commercial contexts, although many hurdles still obstruct their path to wider implementation. The new algorithm seeks to address an issue in efficient load following—a microreactor’s ability to adjust power output based on shifting demand.
Leveraging the physics
For the study, the team ran several simulations for High-Temperature Gas-Cooled Reactors (HTGR), a family of reactors that use helium gas and ceramic materials to stabilize nuclear fission (a process that splits heavy atoms to generate energy). These reactors can range from microreactor sizes to giant, traditional scales, making them good models for the researchers to work with.
The team ran their algorithm on a simplified replica of a microreactor, making sure to preserve key parameters such as power density, coolant temperature, core pressure, and flow velocity. The simulations instructed the algorithm to adjust the power up or down by 20% every minute.
The algorithm stayed within 0.234% of the target measurements for the experiment, without any external help from AI or human controllers. Essentially, “everything about the automated control for load follow operation is grounded in physics and mathematics and readily explainable—an essential feature for passing regulatory review,” the researchers explained. The team conducted additional sensitivity tests by switching up different parameters and found that the algorithm still functioned properly.
The remarkable success of the new algorithm belies the fact that, again, the technology still has a long way to go and that communities haven’t really warmed up to the idea of small nuclear reactors in their communities. That said, it’s a promising result and a reminder of the incredible physics responsible for smoothly operating technological instruments.