A joint LSTM-XGBoost-based method for predicting the state of health and remaining useful life of lithium batteries

The rapid development of the new energy vehicle field poses greater demands on the state of health (SOH) monitoring and remaining useful life (RUL) estimation of lithium batteries for energy storage. In order to further improve the accuracy of SOH and RUL prediction of lithium batteries, a joint prediction method for SOH and RUL of lithium batteries based on long short-term memory (LSTM) and extreme gradient boosting (XGBoost) is proposed. Firstly, incremental capacity analysis (ICA) is used to process the charging stage data and extract the health factors. An LSTM-based joint prediction model for SOH and RUL is then established by leveraging the correlation between the peaks of ICA curves and battery SOH. Subsequently, XGBoost is utilized to optimize the LSTM model, and the error reciprocal method is incorporated to improve prediction accuracy and generalization capability. The model is trained and validated using the National Aeronautics and Space Administration (NASA) dataset. The experimental results show that the combined LSTM-XGBoost prediction algorithm has better prediction accuracy and stability than the single algorithm, with all root mean square error values below 0.02 A·h. This indicates that the LSTM-XGBoost prediction model proposed in this paper shows better performance and application potential in the joint prediction of SOH and RUL of lithium batteries.