CNN-Based Online Diagnosis of Knee-Point in Li-Ion Battery Capacity Fade Curve

Abstract

Accurate monitoring of capacity degradation of a lithium-ion battery is important as it enables the user to manage its usage for optimal performance/lifetime and also to take preemptive actions against any potential explosion or fire. Battery capacity fades gradually through repetitive charging and discharging until it reaches the so called 'knee-point' after which it goes through rapid and irreversible deterioration to reach its end-of-life. It is crucial to forecast the knee-point early and accurately for safe and economic use of the battery. Machine learning based methods have been used to predict the knee-point with early cycle data. Despite some notable progress made, the existing methods make the unrealistic assumption of constant cycle-to-cycle charge/discharge operation. In this study, a deep learning method is developed for online knee-point prediction under the more realistic scenario of variable battery usage. A CNN-based model extracts temporal features of data across past and current cycles to sort out those cells in an urgent state that calls for close monitoring, and then predict the number of cycles left to reach the knee-point. The proposed method extracts features from dynamic data and thus the extracted features reflect dynamic changes in battery properties, thereby improving the prediction performance under realistic scenarios.