Machine learning helps EV batteries to charge in 10 minutes

The researchers presented their results this week at the autumn meeting of the American Chemical Society (ACS).

“Fast charging is the key to increasing consumer confidence and overall adoption of electric vehicles,” said Eric Dufek, Ph.D. “It would allow vehicle charging to be very similar to filling up at a gas station.”

When a lithium-ion battery is being charged, lithium ions migrate from the cathode to the anode. By making the lithium ions migrate faster, the battery is charged more quickly, but the lithium ions do not always fully move into the anode. In this situation, lithium metal can build up, triggering early battery failure. It can also cause the cathode to wear and crack. Added together, these issues reduce the lifetime of the battery and the effective range of the vehicle.

One solution is to tailor the charging protocol in a way that optimises speed while avoiding damage for different types of battery designs, but this requires a significant amount of data on how various methods affect these devices’ lifetimes, efficiencies and safety. The design and condition of batteries, plus the feasibility of applying a given charging protocol with the current electric grid infrastructure, are also key variables.