Accelerated commercial battery electrode-level degradation diagnosis via only 11-point charging segments

Abstract

Accelerated and accurate degradation diagnosis is imperative for the management and reutilization of commercial lithium-ion batteries in the upcoming TWh era. Different from traditional methods, this work proposes a hybrid framework for rapid and accurate degradation diagnosis at the electrode level combining both deep learning, which is used to rapidly and robustly predict polarization-free incremental capacity analysis (ICA) curves in minutes, and physical modeling, which is used to quantitatively reveal the electrode-level degradation modes by decoupling them from the ICA curves. Only measured charging current and voltage signals are used. Results demonstrates that 11 points collected at any starting state-of-charge (SOC) in a minimum of 2.5 ​minutes are sufficient to obtain reliable ICA curves with a mean root mean square error (RMSE) of 0.2774 Ah/V. Accordingly, battery status can be accurately elevated based on their degradation at both macro and electrode levels. Through transfer learning, such a method can also be adapted to different battery chemistries, indicating the enticing potential for wide applications.

Keywords

• Lithium-ion battery
• Degradation diagnosis
• Deep learning
• Physical modeling
• Electrode level