State of health estimation for lithium-ion battery based on improved support vector regression

Abstract The lithium-ion battery has grown to be one of the most popular types energy storage because its many great qualities. However, state health (SOH) decreases as number cycles increases, resulting in reduced performance or even failure device, so an accurate SOH is essential for safe operation device. To solve issue existing estimate methods' low estimation accuracy, a approach based on improved support vector regression proposed. Firstly, changes characteristics during charging and discharging are analyzed factors characterizing degradation extracted. Pearson Spearman correlation coefficients used quantitatively analyze between factors. In addition, by optimizing kernel function parameters SVR model IALO, ant lion optimization (IALO-SVR) technique was developed. IALO-SVR method validated with NASA dataset, experimental results showed that can more accurately compared Back Propagation Neural Network (BP) ALO-SVR methods, error no than 1.8% at most.