Superconducting magnets are commonly used in magnetic resonance imaging (MRI) scanners, which use strong magnetic fields and radio waves to produce detailed images of the inside of the body.
They also have a number of other current and potential future uses, including the next-generation SCMaglev train system in Japan, electric aircraft and nuclear fusion.
One major drawback with the current superconductors is that they utilise coils of superconducting niobium-tin alloy wire that are large, expensive and can take a long time to fabricate.
Now, researchers have used AI to develop a relatively inexpensive and powerful iron-based superconducting magnet, which could open the door to more widespread use of the technology.
Dr Mark Ainslie from King’s College London said that their low-cost and scalable alternative was easier to work with and could lead to smaller and more lightweight devices using superconductors.
He added that, while they still need to be cooled to extremely low temperatures, they could be made more quickly and in smaller devices that could, for example, be installed at a GP’s office rather than a hospital.
MRI machines have very strict requirements regarding the strength and stability of the magnetic field produced by their superconducting magnets.
The researchers claim their new magnets, detailed in the journal NPG Asia Materials, are the first iron-based superconductors to meet these criteria.
Machine learning was used to adjust the fabrication process
The team used a machine learning system called BOXVIA to optimise the creation of superconducting magnets by changing parameters such as heat, and time used in the fabrication process.
The AI can build on initial attempts by the researchers, identifying patterns related to improved performance and fine-tuning the parameters further.
It would usually take months for humans to produce a single magnet and test its properties for performance and behaviour in different situations.
The AI-developed magnets were also found to have different structures at a microscopic level, compared to those developed without the system.
Human researchers have previously tended towards more uniform structures, whereas the AI-designed products featured larger iron-based crystals within the magnet structure.
The team is now set to explore precisely how this never-before-seen nanostructure produces the superior superconducting properties observed.
This could lead to the development of even more powerful magnets in the future.
Today’s news was brought to you by TD SYNNEX – the UK’s number one solutions distributor.
