A State-of-Health Estimation Method of Lithium-ion Batteries using ICA and SVM

Maximum available capacity is often used as the health indicator for the state-of-health estimation of lithium-ion batteries. However, the incomplete charging and discharging process in the field usually is encountered. They are resulting in the maximum available capacity not being measured by the ampere-hour integral method. To improve the state-of-health estimation accuracy, we first propose four new fade features based on incremental capacity analysis. Then support vector machine is utilized to model the non-linear function between features and health index. Finally, a case study is conducted to verify the effectiveness and performance of the proposed method.