A Semi-Empirical State of Health Estimation Method for Batteries of Electric Vehicles Operating in Varying Real-World Conditions

Present research on battery State of Health (SoH) estimation is mainly focusing on constant conditions, solely cycling conditions or solely storage conditions, whereas few attempts have been made for varying real-world conditions. Since batteries of electric vehicles (EVs) usually operate in varying conditions, it is of great importance to break the constraints of constant conditions, solely storage conditions and solely cycling conditions, while estimating SoH for EV batteries. In light of this, a semi-empirical SoH estimation method for EV batteries operating in varying conditions is proposed in this paper, which includes the offline parameterization and the real-world calculation. Firstly, the offline parameterization is conducted by fitting two semi-empirical models to the separate storage degradation data and the cycling degradation data from experiments and by building two parameter look-up tables for storage conditions and cycling conditions respectively. Subsequently, the real-world calculation is implemented by using the look-up tables in combination with real-world varying conditions and by taking the real-world alternation between storage conditions and cycling conditions into account. The proposed method is not only capable of estimating SoH with a good accuracy for EV batteries operating in varying conditions but also can quantify the contributions to the overall SoH variation of the EVs’ different types of conditions, such as parking, charging, driving (discharging) and regenerative braking. A Jaguar I-PACE EV was taken as the test vehicle and the results show that the estimated SoH was approximately 1% different from the tested SoH.