State of health estimation for lithium-ion battery via charging time for partial voltage range

The state of health (SOH) estimation for lithiumion battery is necessary to ensure the reliability and safety of electric vehicles. However, the SOH is related to complex chemical reactions and coupled with multiple physical quantities, it exhibits non-linear characteristics. In this paper, a method based on support vector regression (SVR) and back propagation neural network (BPNN) is proposed to estimate the health state of the battery when the battery is not fully charged and discharged. The length of the charging time in a partial voltage range is selected as the health index. Firstly, the current and voltage data of the battery were obtained by aging cycle test under constant current and constant voltage schedule. Secondly, using Gaussian filter to obtain a smooth IC curve and determine the voltage range where the curve changes dramatically. Thirdly, taking the charging time in the above voltage range as HF and the input of models. The voltage range selected is easily accessible in the pratical application. Results demonstrate that the proposed method provides an accurate SOH estimation.