Argonne’s machine learning model is equipped to analyze data from 31 sensors at its Mechanisms Engineering Test Loop (METL) facility.
Machine learning technology has the potential to transform nuclear reactor operations, according to a team of experts from the US Department of Energy's Argonne National Laboratory, who demonstrated how it may improve security and efficiency.
They showcased the application of machine learning in the sodium-cooled fast reactor (SFR), a specialized cutting-edge nuclear reactor.
It is a type of nuclear reactor that employs liquid sodium as a coolant for its core. This application enables it to efficiently generate electricity without producing carbon emissions through splitting heavy atoms.
While these reactors are not currently utilized for commercial purposes in the United States, there is widespread optimism that they hold the potential to revolutionize power generation and contribute to the reduction of nuclear waste. In the near future, SFRs are viewed as a possible path for cleaner and more sustainable energy generation.
See Also
Related
- Machine learning might help us finally unlock nuclear fusion
- Nobel laureate to build rapid-fire laser-powered nuclear fusion reactor by 2030
- A special new tape could make small, efficient nuclear fusion reactors possible
The need for an ML system
The official release highlighted that currently, the technology encounters a significant hurdle related to ensuring the “purity of their high-temperature liquid sodium coolant.”
Maintaining this purity is critical to preventing corrosion and system clogs. In response to these issues, Argonne scientists created this game-changing machine learning (ML) system.
“By harnessing the power of machine learning to continuously monitor and detect anomalies advances the state of the art in instrumentation control. This will create a breakthrough in the efficiency and cost-effectiveness of nuclear energy systems,” said Alexander Heifetz, principal nuclear engineer at Argonne, in the press statement.
Here's what all the ML model can do
The team created a machine-learning model while taking into account a variety of operational criteria.
The first was for the ML model to be capable of continually monitoring the cooling system.
“The model is equipped to analyze data from 31 sensors at Argonne’s Mechanisms Engineering Test Loop (METL) facility that measure variables like fluid temperatures, pressures, and flow rates,” added the release.
The METL facility is a one-of-a-kind experimental setup developed to assess materials and components proposed for use in these reactors in a safe and precise manner.
It also serves as a training ground for engineers, technicians, and even machine learning models, all of whom could contribute to the operation and maintenance of these reactors. The incorporation of a complete system augmented by machine learning might improve monitoring, lowering the danger of abnormalities that could disrupt reactor operation.
The team also demonstrated the model's capacity to detect operational irregularities quickly and correctly.
“They put this to the test by simulating a loss-of-coolant type anomaly, which is marked by a sudden spike in temperature and flow rate. The model detected the anomaly within approximately three minutes of its initiation,” explained the release.
