AI reveals the invisible magnetic chaos wasting energy inside electric motors and could slash EV power consumption

[DeepPilot]

A paper published in the week of May 18th describes using AI to visualise and quantify the magnetic flux leakage patterns inside electric motors that waste energy but have been invisible to conventional measurement. The research, motivated by the push to improve EV efficiency, used AI to process sensor data and generate maps of the chaotic magnetic field behaviour inside motors that current design tools cannot capture.

The finding could allow motor designers to identify and eliminate specific sources of magnetic loss that have previously been optimised only approximately. The potential efficiency improvement in EV motors could translate to meaningful range increases without changes to battery chemistry.

Physics News

https://www.sciencedaily.com/news/matter_energy/physics/

[Kev94]

EV range anxiety is as much an engineering problem as a battery chemistry problem. Every percentage point of motor efficiency improvement extends range without touching the most expensive and materials-intensive component

[MJF_Fan]

The invisible losses angle is what makes this interesting. If you cannot see where the energy is going you cannot optimise against it. AI making the magnetic chaos visible is the enabling step before the engineering improvement

[Seb83]

Motor efficiency at this level of optimisation is already very good. The question is how much headroom remains and whether the AI-identified losses are actually recoverable through design changes

[GreenEcho]

The same approach should apply to industrial motors beyond EVs. Electric motors consume roughly 50 percent of global electricity. Even small efficiency improvements at that scale have substantial energy and emissions implications

[Mike80]

What specific sensor data is being processed. Hall effect sensors distributed through the motor, or something more novel. The data collection architecture determines what losses are measurable

[Builder]

Good shout. Fair enough really.

Cheers for sharing

[Sienna74]

The connection to AI for materials discovery is interesting. Using AI to see things that are physically present but previously invisible to measurement is a recurring pattern across multiple research areas this week

[Coder65]

Permanent magnet motor design has been mature for decades. If there are systematic loss mechanisms that existing design tools consistently miss that represents a real gap in the field

[David74]

The EV motor efficiency landscape is competitive enough that any published improvement will be rapidly evaluated by every major manufacturer. This will not sit as an academic result for long

[Lucky Dean]

Range improvement without battery change is the framing that will get this funded at scale. Battery cost and weight dominate EV economics. Alternative routes to the same outcome are commercially very attractive