State-of-health estimation for lithium-ion batteries by combining model-based incremental capacity analysis with support vector regression

Abstract

Accurate state-of-health (SOH) estimation for lithium-ion batteries is of great significance for future intelligent battery management systems. This study proposes a novel method combining voltage-capacity (VC)-model-based incremental capacity analysis (ICA) with support vector regression (SVR) for battery SOH estimation. For accurate and efficient capture of IC curves, 18 VC models are first compared, and then, suitable models are selected for two types of batteries with different chemistries, enabling multitype health features to be obtained by parameterizing the VC models. After correlation analysis of these extracted health features with the reference battery capacity, the SVR algorithm is adopted to construct SOH estimation models. Finally, four aging datasets are employed for validation of the proposed method. The experimental results show that the SVR models achieve high accuracy in SOH estimation, i.e., the respective mean absolute errors (MAEs) and root mean square errors (RMSEs) of all batteries are limited to within 1.1%. Moreover, the method is robust against different initial aging statuses and cycle conditions of the batteries: after migration and fine-tuning, both the MAEs and RMSEs can be confined to within 2.3% by utilizing the established SVR models.