Electrochem Seminar: "High-throughput experiment and modeling to predict lithium plating during fast charging"
Electrochem Seminar: "High-throughput experiment and modeling to predict lithium plating during fast charging"
Abstract
Fast charging of most commercial lithium-ion batteries is limited due to fear of lithium plating on the graphite anode, which is difficult to detect and poses significant safety risk. We previously used high-throughput cycling techniques to quantify irreversible Li plating and understand the effects of energy density, charge rate, temperature, and State-of-Charge (SOC) on plating from over 70 Lithium|Graphite half-cells. The results were used to refine a mature physics-based electrochemical (EChem) Newman model as well as create a simple and interpretable empirical equation for predicting the plating onset SOC.
Now, we aim to address the real-world challenge of predicting lithium plating for cells with erratic temperature and SOC, which are values not directly measured by battery control systems. This talk will introduce a data-driven lithium plating model based on cell voltage change throughout charge, built using synthetic data from the previously calibrated EChem model. We will discuss data visualization methods, feature engineering strategies, preliminary model performance, and physical insights gained. Overall, we highlight the workflow from experiment to reliable physics-based models and then creating efficient data-driven models from physics-based synthetic data.