Data-driven SOH Estimation of Lithium-ion Batteries Based on Savitzky-Golay Filtering and SSA-SVR Model

State of health (SOH) estimation is the core problem of the lithium-ion battery system. Given the low estimation accuracy and weak practicability of traditional methods, this paper proposes a SOH estimation method based on sparrow search algorithm and support vector regression (SSA-SVR). Firstly, four features are extracted from the voltage and current curve segments during the charging process. Extracting these features only requires segment data, which has high extraction efficiency and strong practicability. Secondly, aiming at the characteristics of the current sensor that is easily affected by noise, Savitzky-Golay filtering is performed on the two features extracted from the current data to ensure the stability of the model. Finally, SSA is used to optimize SVR parameters, and the SOH estimation model is constructed. The proposed model is verified on the NASA battery data set, and the results show that it has high estimation accuracy, strong applicability, and good robustness.