Sensors-enabled approach for real-time quantification of lithium plating under extreme environments

Ensuring safe and efficient fast charging of lithium-ion batteries (LiBs) in low-temperature environments remains challenging due to lithium plating on the anode under extreme conditions, which compromises battery safety and longevity. In this paper, we introduce advanced sensors into a LiB that integrates state-of-the-art multi-dimensional sensing technologies for real-time, in-situ detection and quantification of lithium plating. This innovation achieves unparalleled functionality without altering battery's physical dimensions. It provides dynamic, high-resolution insights into internal pressure, temperature, and anode potential, enabling the extraction of multi-dimensional features closely linked to lithium plating. By leveraging advanced statistical approaches, including correlation analysis and least absolute shrinkage and selection operator regression, the critical features are identified and ranked. These features are further integrated using a cutting-edge machine learning framework combining feature distance-based analysis with Adaboost. Only six features during battery charging are required as input, the model achieves remarkable lithium plating quantification accuracy of 93.3 % at a single temperature and 88.5 % at different temperatures. This sensors-enable approach to lithium plating quantification offers a promising pathway toward enhancing the functionality and intelligence of next-generation battery management systems for electric vehicles and portable electronic devices.