A Novel State of Health Estimation Method for Lithium-Ion Batteries Based on Incremental Capacity Curve

As the use of electric vehicles (EVs) continues to grow, the safety of their battery systems has raised significant concerns. Consequently, a deep understanding of battery degradation characteristics, along with the development of models for accurately predicting the state of health (SOH), has become a critical challenge hindering the further growth of this technology. Incremental capacity analysis (ICA) is a widely employed method for SOH estimation, with the determination of the incremental capacity (IC) curve being a key step. However, various challenges related to low sampling frequency and incomplete charging hinder the accurate determination of the IC curve. To effectively address these issues, a novel SOH prediction method for lithium-ion batteries (LIBs) based on the IC curve is proposed in this work. From our experimental results on LIBs during battery degradation, it was demonstrated that the position of peak A on the IC curve is strongly correlated with SOH. The IC curve features obtained using Akima nonlinear interpolation and Lowess smoothing (AILF) show superior correlation with SOH compared to traditional methods. The method demonstrates high prediction accuracy across battery life cycles. The proposed method was validated using the publicly available National Aeronautics and Space Administration (NASA) datasets, achieving RMSE values of 1.38% for the B0007 battery and 1.50% for the B0018 battery, demonstrating superior accuracy compared to existing methods.