Researchers from the University of Central Florida, including Shanthan Kumar Padisala, Bharatkumar Hegde, Ibrahim Haskara, and Satadru Dey, have developed a novel approach to optimize battery charging protocols that adapt to battery degradation over time. Their work, published in the Journal of The Electrochemical Society, combines physics-based modeling with reinforcement learning to enhance the widely used Constant Current Constant Voltage (CCCV) charging protocol.
Batteries degrade with use, primarily through increased resistance and reduced capacity. While various charging methods have been proposed to optimize charging speed and battery health, few have adapted the CCCV protocol to account for this degradation. The researchers aimed to fill this gap by developing a framework that adjusts the constant current part of the CCCV protocol based on the battery’s changing health.
The team used a physics-informed reinforcement learning (RL) approach, which estimates a key degradation mechanism—Loss of Active Material (LAM)—and adjusts the charging current accordingly. They combined PyBamm, an open-source battery modeling tool, with Stable-baselines, a reinforcement learning library, to train the RL agent using a Proximal Policy Optimization (PPO) network.
Simulation results demonstrated the potential of this framework to enhance the CCCV protocol by embedding physics-based information into the RL algorithm. The researchers compared their proposed agent with two other charging protocols: one generated by a non-physics-based RL agent and another using a constant CCCV for all cycles. The adaptive approach showed promise for improving battery longevity and performance.
For the energy sector, this research offers a practical application in extending battery life and optimizing charging processes. By adapting charging protocols to the real-time health of batteries, energy storage systems can operate more efficiently and cost-effectively. This is particularly relevant for large-scale energy storage applications, such as grid storage and electric vehicle fleets, where maximizing battery lifespan is crucial for reducing operational costs and environmental impact.
This article is based on research available at arXiv.