Researchers at Chalmers University of Technology in Sweden have developed an artificial intelligence-based charging method that could extend the life of electric vehicle batteries without increasing fast-charging times. The method adapts the charging current during each fast-charging session according to the battery’s chemistry and state of health. In a recently published study, the researchers found that the approach increased battery life by almost 23 per cent compared with standard charging methods, while keeping charging time largely unchanged.
Fast charging is an important factor for electric vehicle use, particularly for long-distance driving, commercial vehicles, taxis and industrial transport. However, it can accelerate battery degradation because high currents increase the risk of chemical side reactions inside battery cells.
One of the key risks associated with fast charging is lithium plating, a process in which metallic lithium forms on an electrode rather than being stored properly in the battery’s structure. This can reduce battery capacity and, in some cases, affect safety. The researchers said the risk increases as batteries age, while many current charging methods apply the same current and voltage regardless of battery condition.
Changfu Zou, professor at the Department of Electrical Engineering at Chalmers, conducted the study with Meng Yuan, assistant professor at Victoria University of Wellington in New Zealand and a former Chalmers researcher. The study, Lifelong Reinforcement Learning for Health-Aware Fast Charging of Lithium-Ion Batteries, was published in IEEE Transactions on Transportation Electrification.
The AI-based strategy uses reinforcement learning, a machine learning method in which an algorithm improves its decisions through feedback. The researchers trained the model in a simulated environment based on one of the most common types of electric vehicle batteries. The simulation included parameters affecting both charging time and battery health.
The model was trained to adjust charging according to the battery’s charge level and overall health. According to the researchers, the resulting strategy reduced long-term degradation while maintaining charging speed.
“We show that it is possible to charge more or less as fast as today, but with significantly less long-term degradation of the battery,” Yuan said.
The researchers said the method could in principle be introduced through software updates to vehicle battery management systems. However, calibration would be required before it could be applied broadly across battery types.
“There are not so many different battery types today, but the method needs to be calibrated for it to be used by everyone,” Zou said. “Using transfer learning, we can take advantage of what our AI model has already learned, and thus adapt the AI model to new batteries more quickly.”
The next stage of the research is to test the charging strategy on physical batteries. The researchers said longer battery life could support wider use of electric vehicles and reduce costs associated with warranties, resale values and the use of battery raw materials.
