Cross-Cut IV

Data-driven methods leverage high-fidelity experimental data, such as temperature profiles, voltage-capacity curves, and impedance spectroscopy, to train predictive models. These models help identify key electrochemical signatures such as lithium plating, growth of the SEI, and thermal runaway, that are directly linked to degradation pathways. For instance, as shown in the image, experimental data on lithium deposition is combined with predictions from physics-based simulations to detect early signs of degradation, ensuring that corrective actions can be taken before catastrophic failure occurs.

Safety is a key concern in battery systems, especially under operational extremes such as fast charging, over-discharge, or high temperatures. Data-driven methods combined with thermal analytics, including calorimetry, and sensing, allow for accurate prediction of thermal behavior in batteries. The integration of thermal data with physics-based models allows for the identification of critical thresholds where thermal instabilities might occur. By training models on these thermal signatures, the analytics platform can predict potential thermal runaway events and mitigate them through optimized thermal management strategies.

At CARES, we aim to develop data-driven methods that can revolutionize battery management that can offer real-time electrochemical and thermal insights, predictive analytics, and advanced thermal management strategies. By leveraging the synergy between physics-based modeling, experimental data, and machine learning algorithms, the analytics platform provides a comprehensive toolkit to enhance the safety, lifetime, and performance of next-generation battery systems.