Researchers at Chalmers University of Technology in Sweden have developed an artificial intelligence method that increases electric vehicle battery life by nearly 23 percent. The system adapts the fast-charging process to the specific health and chemistry of the battery without increasing the total charging time. 

Implementing this change requires only an update to the vehicle’s battery management software, which makes it a potentially accessible solution for existing electric vehicles.

Currently, electric vehicle batteries have an estimated life of eight to 15 years depending on usage. The availability of fast charging is a significant factor for consumers and companies, especially for operators of heavy industrial vehicles, taxis, and motorists driving long distances. 

While necessary for these applications, frequent fast charging is known to stress battery cells and accelerate the degradation process over time.

AI-driven adaptation via reinforcement learning

Existing charging standards use the same current and voltage levels regardless of whether a battery is new or several years old. This lack of adaptation increases the risk of lithium plating, a process where metallic lithium precipitates on the electrode instead of being stored correctly. 

“This can reduce capacity and may also affect safety, as unevenness in the structure of the lithium can, in a worst case scenario, cause a short circuit,” said the researchers in a press release.

The researchers found that their AI method maintains charging times within a few seconds of current standard speeds while reducing this internal wear.

To address these issues, Professor Changfu Zou and Assistant Professor Meng Yuan developed a strategy based on reinforcement learning. The AI was trained using a digital model of a common electric vehicle battery and simulations of variables that impact both health and charging speed. 

“The AI model was trained to adapt the charging according to how charged or discharged the battery was at the time of charging,” explained the researchers.

“It also needed to take into account the overall health of the battery, as this is crucial to both capacity and electrochemistry. The result was a charging strategy that both keeps the charging time short and minimizes harmful reactions.”

A cost-effective strategy

The study demonstrates that it is possible to maintain current charging speeds while reducing long-term degradation. The researchers state that the strategy is cost-effective to deploy because it works through existing battery management hardware. 

“Our study shows that smart adaptation of the current during charging, taking into account the changing electrochemical state of the battery, can maximize both the performance and the life of the battery,” noted Changfu Zou, professor at the Department of Electrical Engineering at Chalmers.

While the method requires calibration for different battery types, the team intends to use transfer learning to adapt the AI model to new battery chemistries more quickly.

“The next step is to test the method directly on physical batteries,” concluded the press release. “The researchers hope that the AI-based charging strategy will make a significant contribution to the electrification of the transport sector.”