Physics, not AI, could supply these self -regional microreactors

In our world of algorithms dominated by AI, some nuclear physicists prefer to do old science. And this algorithm without AI, without AI, can be the breakthrough necessary for small nuclear reactors to take off seriously.

A recent article published in progress in nuclear energy describes an algorithm that allows nuclear micro -reaches – transportable, mini reactors providing energy to distant locations – to automatically adjust their power. This is particularly relevant for micro -rector, because their placement provided in rural communities, disaster areas and cargo cargo means that there will not always be any personnel available for manual examinations, unlike conventional reactors.

“Our method can help suppliers design reactors with autonomous control systems that are safer and safer,” said Brendan Kochunas, main study and nuclear engineer at the University of Michigan, in a press release.

“Many startups companies and inherited in the United States are pushing towards a short-term and wide deployment of nuclear micro-rector,” he added. “And our work establishes a clear path to achieve it in an economically viable way.”

Small but powerful

Indeed, microreactors are a powerful but mobile option to provide electricity to distant areas. To be clear, micro -reaches are different from small modular reactors, which are reduced versions of conventional nuclear reactors. Microreactors have a smaller energy capacity that generates up to 20 thermal energy megawatts.

In the United States, microreactors exist in the middle of the 20th century, mainly for military use. With the progress of nuclear physics, the micro -reacchers have started to appear in more commercial contexts, although many obstacles always obstruct their path to a wider implementation. The new algorithm seeks to solve a problem in the following effective load: the capacity of a micro -rector to adjust power as a function of exchange demand.

Take advantage of physics

For the study, the team organized several simulations for reactors cooled by high temperature gas (HTGR), a family of reactors that use helium gases and ceramic materials to stabilize nuclear fission (a process that divides heavy atoms to generate energy). These reactors can range from the size of micro -reaches to giant and traditional scales, which in fact good models for researchers with researchers.

The team has executed its algorithm on a simplified replica of a micro -rector, making sure to preserve key parameters such as power density, cooling liquid temperature, central pressure and flow speed. The simulations have asked the algorithm to adjust power up or the 20% decrease every minute.

The algorithm has remained at less than 0.234% of the target measures for experience, without any external AI assistance or human controllers. Essentially, “everything on automated control for the operation of the load monitoring is based on physics and mathematics and easily explained – an essential characteristic for passing a regulatory review,” explained the researchers. The team carried out additional sensitivity tests by changing different parameters and found that the algorithm was still working properly.

The remarkable success of the new algorithm denies the fact that, once again, technology still has a long way to go and that the communities have not really warmed up at the idea of ​​small nuclear reactors in their communities. That said, it is a promising result and a reminder of the incredible physics responsible for gentle technological instruments.