Remaining useful life prediction method of EV power battery for DC fast charging condition

Abstract

To realize the online rapid prediction of the remaining useful life (RUL) of electric vehicle (EV) power battery under direct current (DC) fast charging conditions and reduce the influence of complex health indicator (HI) on the prediction accuracy, an adaptive prediction method of the RUL of EV power battery based on whale swarm algorithm–long short-term memory (WSA-LSTM) algorithm is proposed in this paper. Firstly, a complete set of HI is built based on a dynamic data-driven application system (DDDAS) to ensure the real-time combination of condition monitoring data and simulation systems. Then, combined with the calculation of Pearson and Spearman correlation coefficients and the entropy weight method, the main features that affect the battery capacity are screened out to reduce the interference of secondary HI on RUL prediction. Thirdly, the whale swarm algorithm is used to globally optimize the hyperparameters of the long short-term memory to achieve rapid RUL prediction. Finally, we conduct simulation experiments based on real-time data under EV DC charging conditions. The simulation results show that the HI screening method proposed in this paper is effective, and the RUL prediction method can achieve a fast and accurate prediction of DC fast charging conditions.

Keywords

• DC charging condition
• Lithium battery
• Health indicator
• Remaining useful life