Chalmers Research Shows AI can Extend EV Battery Life by 23%
EVs have gained traction partly because they eliminate issues common to petrol and diesel engines.
However, battery degradation now poses a maintenance concern for the sector. Lithium-ion batteries deteriorate with frequent fast charging.
A study now shows that AI can slow this degradation and extend the battery life of EVs.
Researchers at Chalmers University of Technology developed an AI-based charging method developed.
The study, published by academic journal IEEE, reveals that the method can extend the vehicle’s battery life by as much as 23%.
AI optimises charging currents
The team at Chalmers developed a system that adjusts current during fast-charging cycles.
Meng Yuan, a postdoctoral researcher at the university's Department of Electrical Engineering, says: "This work introduces the first explicit formulation of a lifelong battery fast charging problem.
"The proposed method achieves a significant improvement in performance, where battery lifespan is extended to 703 equivalent full cycles… representing a 22.9% improvement over the standard baseline."
Modern EV batteries last for years under normal conditions. Fast charging can accelerate aging by stressing cell components and causing lithium plating, where ions accumulate on the anode.
The battery management system uses reinforcement learning to prevent this degradation.
This machine-learning technique allows systems to identify optimal outcomes through trial and error. The technology continuously learns from each charging session, refining its approach to protect battery health while maintaining the rapid charging speeds that drivers expect from modern charging infrastructure.
Adapting to battery condition
The Chalmers team applied reinforcement learning to adjust current based on pack chemistry and condition during fast charging.
As batteries age, the AI modifies voltage to reduce stress on the anode, cathode and electrolyte.
The system monitors multiple parameters including temperature, voltage and charge state to make real-time decisions. This dynamic approach differs from conventional charging protocols that apply fixed parameters regardless of individual battery condition or environmental factors.
According to a 2024 study by Geotab, average annual EV battery degradation sits at around 1.8% per year. This rate suggests batteries could last at least 20 years or 200,000 miles, with many exceeding this range.
Tesla batteries can last from 300,000 to 500,000 miles, depending on usage and charging patterns, according to some estimates. A 23% improvement could translate to nearly 70,000 extra miles on the low end and more than 100,000 additional miles on the high end.
The researchers measured battery life extension through charge and discharge cycles.
The study authors say: "The proposed approach maintains comparable charging efficiency while largely extending battery lifespan, demonstrating that lifespan enhancement can be achieved without compromising charging speed".
Real-world testing pending
According to the Federal Highway Administration, Americans drive about 13,476 miles each year on average.
For drivers using fast charging frequently, this AI-enabled method could extend vehicle ownership by several years.
The charging experiment took place in a laboratory and not on physical batteries. Real-world validation remains necessary before the technology can affect battery warranties and the used EV market.
The research provides a framework for applying AI software to hardware challenges. The method could inform how manufacturers approach battery management systems in future vehicle models.
