Energy Storage Science and Technology ›› 2025, Vol. 14 ›› Issue (1): 331-345.doi: 10.19799/j.cnki.2095-4239.2024.0675

• Energy Storage Test: Methods and Evaluation • Previous Articles     Next Articles

Remaining useful life prediction of lithium-ion battery based on an ABC-LSTM model

Yong LIU1(), Huaiwen YU2(), Dapeng LIU1, Yong MU1, Yingzhou WANG2, Xiuyu ZHANG2   

  1. 1.State Grid Jibei Electric Power Co. , Ltd. Tangshan Power Supply Company, Tangshan 130033, Hebei, China
    2.School of Automation Engineering, Northeast Electric Power University, Jilin 132012, Jilin, China
  • Received:2024-07-22 Revised:2024-08-10 Online:2025-01-28 Published:2025-02-25
  • Contact: Huaiwen YU E-mail:liu.y.e@jibei.sgcc.com.cn;2202300730@neepu.edu.cn

Abstract:

To ensure the safe and stable operation of energy storage systems, the accurate prediction of the remaining useful life (RUL) of lithium-ion batteries is crucial. This study presents an integrated forecasting model that combines the artificial bee colony (ABC) algorithm with a long short-term memory (LSTM) network enhanced by dropout techniques. This combination effectively improves the accuracy of RUL predictions for lithium-ion batteries. First, the dropout regularization method is utilized to effectively mitigate overfitting, thereby enhancing the generalization capability of the predictive model. Subsequently, an activation layer network structure is introduced to address capacity recovery and data noise issues, significantly enhancing the ability of the model to handle complex nonlinear data. Thereafter, the hyperparameters of the LSTM-based comprehensive forecasting model are optimized using the ABC algorithm to avoid local optima and improve the precision of RUL predictions. Finally, the predictive accuracy and robustness of the proposed model are verified using a public dataset from the NASA Research Center and the CALCE. The paper conducts an experimental analysis and verification of The predictive performance of various algorithms were experimentally analyzed and verified using training data at 40% and 60% levels. The performance of swarm optimization algorithms, such as the Sparrow Search Algorithm and the Humpback Whale Optimization Algorithm, were also compared. The experimental results demonstrate that the proposed ABC-LSTM integrated forecasting model can capture the global trends and local characteristics of the capacity degradation of the lithium-ion battery more accurately than the compared models. The root-mean-squared error of the RUL prediction results obtained with a 60% proportion of training data remained consistently within 1.02%; the mean absolute error remained consistently within 0.86%, and the fitting coefficient exceeded 97%.

Key words: lithium-ion battery, remaining useful life prediction, long short-term memory network, artificial bee colony algorithm, dropout technology

CLC Number: