A practical state of health estimation method for lithium-ion batteries using charging duration in different voltage segments

Accurate state of health (SOH) estimation for lithium-ion batteries is essential for the safety and reliability of electric vehicles and energy storage system. However, the variability and unpredictability of actual charging voltage ranges limit the applicability of many existing methods. This study estimates SOH using a data-driven method that utilizes the charging duration of different voltage segments, making it applicable to a wide range of charging voltages. The whale optimization algorithm-optimized radial basis function neural network (WOA-RBFNN) is employed for SOH estimation. Five voltage segments with significantly changes due to battery aging are selected based on incremental capacity analysis (ICA), instead of directly selecting the charging duration of the voltage segments or the peaks and valleys from the incremental capacity curve (ICC). It is verified through ICA that the charging duration is proportional to the ICC’ envelope area, making it a suitable health indicator (HI) for each voltage segment. Then, the WOA-RBFNN is employed to construct the mapping between charging duration and SOH. The proposed method is validated using two representative datasets. The results demonstrate that the mean absolute error is below 0.589%, and the root mean square error is below 0.701%. The proposed method is applicable to a wide range of charging voltage, enhancing its practical usability.