Degradation Modes Identification of Lithium-Ion Batteries Based on Flexible Discharge Data

The precise identification of battery degradation modes stands as a critical aspect for the enhancement of battery system management, augmenting system safety, and fortifying overall system stability. However, the complex and variable nature of electrochemical processes, alongside the constraints posed by current testing apparatus, presents formidable obstacles to the real-time assessment of battery aging. In this study, we initially established the profound linkage between discharge curves and battery aging through an array of battery aging experiments. A meticulous examination was conducted on the correlation between segments of the discharge curve within specific voltage intervals and battery aging. This was followed by the accurate determination of battery aging patterns through the utilization of flexible discharge curve segments, integrated within a deep learning framework. Furthermore, we shed light on the critical importance of potential data encapsulated in discharge curves for depicting battery aging, corroborated by visual demonstrations of our model's intermediate processes. Our findings highlight the significant promise of employing readily available battery data for unlocking deeper understanding of the dynamics driving battery aging.