State-of-Health Estimation Under Multistage Constant Current Charging Pattern Considering Future Charging Current Prediction

The multi-stage constant current charging (MSCCC) pattern can effectively enhance charging efficiency and even ensure thermal safety. However, current switching in this charging pattern causes voltage fluctuation, leading to outliers in incremental capacity curve construction and affecting aging features construction. Besides, it is necessary to determine more voltage intervals under partial charging conditions to construct aging features, providing more input information conducive to data-driven-based State of Health (SOH) estimation modeling. This work first proposes a multi-stage voltage interval determination strategy for constructing effective aging features while considering partial charging. Additionally, an online updating strategy of SOH estimation model based on multi-state information is proposed, so that the established model can be quickly adapted to different battery applications. Finally, a future charge current prediction strategy is designed to satisfy the construction of aging features and SOH estimation in practical applications. Aging experimental results confirm the effectiveness of the proposed strategies for constructing aging features and the benefits of the multi-state aging features for SOH estimation. The designed future charging current prediction strategy enables accurate expected charging curves.