Li-ion Battery SOH Estimation Model Based on Finite Sampling in Relaxation Phase

Accurately estimating the state of health of Li-ion battery is crucial for the safe operation of battery energy system. The effectiveness of features can be influenced by temperature variations and nonzero State of Charge. However, the relaxation phase after fully charged is relatively unaffected by the charging process. Therefore, this study proposed constructing relevant geometric features by extracting a finite number of voltage sampling points from the relaxation phase. Subsequently, the Pearson correlation analysis is performed on the derived geometric features, followed by the application of Principal Component Analysis to further reduce feature dimensionality, eliminating redundant information and decreasing the overall complexity of the model. Finally, the effectiveness of the proposed features is validated using Kernel Ridge Regression and Support Vector Regression models for NCA and NCM batteries across diverse temperature cycling conditions. The results indicate that the proposed SOH estimation model maintains a high level of accuracy and robustness.