AI-Managed EV Charging May Increase Battery Life by Approximately 23%

      CORTE MADERA, CALIFORNIA - FEBRUARY 15: An aerial shot shows Tesla vehicles recharging at a Tesla charging station on February 15, 2023, in Corte Madera, California. The electric vehicle manufacturer Tesla is collaborating with the U.S. federal government to enhance electric vehicle charging infrastructure across the nation. Tesla has announced intentions to open approximately 7,500 of its Supercharger stations to all electric vehicle brands by the conclusion of 2024.

      Justin Sullivan/Getty Images

      The most significant automotive news and reviews, without the fluff

      Our daily newsletter offers crucial stories delivered to your inbox every weekday.

      As car manufacturers strive for more potent EV charging solutions to provide quicker charge times, battery packs may reveal vulnerabilities. Frequent DC fast charging can negatively affect the long-term health of electric vehicle batteries, and older batteries are generally less capable of enduring intensive charging. In light of this, researchers have explored whether AI could offer a solution.

      In a recent paper published by IEEE, researchers at Chalmers University of Technology created a “health-aware” charging algorithm that can assess a battery’s state of health and adapt its charging behavior as the battery ages. According to the researchers, simulations indicated that this approach could reduce projected degradation, extending a battery’s usable life by about 23% while maintaining charge times comparable to new batteries.

      The AI-driven algorithm can interface with a battery monitoring system, “learning” its health status over time and identifying potential issues. Researchers suggest that this approach is effective enough to eliminate the necessity for dedicated sensors to monitor the battery itself. While most automakers currently assess battery health by monitoring voltage at the cell level, the researchers propose that examining the chemical reactions within the cell might be essential for achieving similar results.

      Depending on the battery's health, the algorithm can adjust the charging speed by configuring different voltage limits. The paper states that with this method, a simulated battery successfully completed 703 charge and discharge cycles before its capacity fell below 80%, whereas a simulated battery charged at a constant voltage achieved only 572 cycles. Charging times remained nearly identical, with the traditional method taking 24.15 minutes and the AI method taking 24.12 minutes.

      Although promising research initiatives do not always result in commercially viable products, software capable of monitoring battery health in real time and adjusting charging accordingly is progressing toward market readiness—regardless of whether it carries the “AI” label. At CES that year, GBatteries introduced its own charging software, while Breathe originated from an Imperial College London research project. Breathe is now collaborating with Volvo to offer charging software for the automaker's upcoming generation of electric vehicles, beginning with the 2027 EX60. Besides enhancing health, Breathe claims it can boost charging speed by as much as 15-30% through real-time data monitoring.