Chalmers researchers develop AI charging protocol extending EV battery lifespan by 23%.
Most battery degradation occurs during suboptimal fast-charging cycles that stress cell chemistry. By using AI to dynamically optimize charging currents in real-time, this research offers a software-level fix to a hardware bottleneck. If commercialized, it could significantly lower EV total cost of ownership without requiring new battery chemistries.
What Happened
Researchers at Chalmers University of Technology have developed an AI-driven charging model that extends the lifespan of electric vehicle (EV) batteries by up to 23%. Notably, this improvement in cycle life is achieved without extending the overall charging time, presenting a highly efficient software-based optimization for existing battery infrastructure.Technical Details
Traditional EV fast-charging protocols typically rely on static, predefined current profiles, such as Constant Current-Constant Voltage (CC-CV). These rigid profiles often fail to account for the dynamic, real-time electrochemical states of individual battery cells, leading to localized stress, lithium plating, and accelerated degradation.The Chalmers team utilized machine learning algorithms to create a dynamic charging protocol. The AI model continuously analyzes battery health parameters, dynamically adjusting the charging current to minimize stress on the battery's internal chemistry. By optimizing the charge delivery at a granular level, the system mitigates the micro-degradation events that cumulatively shorten battery life, achieving optimal charge rates without the penalty of thermal or chemical damage.