Scholarlink

<p>The capacity prediction of lithium battery can guide the replacement of batteries that are about to be retired, which is the most important step of battery management system (BMS). In order to predict the capacity degradation under different conditions, this paper proposes a novel method using long short-term memory (LSTM), particle swarm optimization (PSO), and singular spectrum analysis (SSA). Firstly, SSA decomposes the raw capacity degradation trend into global degraded trend, reversible degradation, and noise. Then, LSTM is adopted to learn and predict the capacity degradation. Finally, PSO automatically optimizes the initial learning rate and the number of hidden layers of LSTM for improving the prediction accuracy of battery capacity degradation. The proposed SSA-PSO-LSTM method is validated by different lithium-ion batteries from the NASA dataset. The experimental results show that both the root mean square error (RMSE) and the mean absolute error (MAE) of the prediction results are less than 1%, demonstrating the high accuracy and effectiveness of the proposed approach.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Battery capacity degradation prediction based on singular spectrum analysis and improved deep learning

  • 期刊论文

  • 发表日期：

    2025年7月31日

查看全文
  • Xingbo Zhang,
  • Kui Chen,
  • Yang Luo,
  • Xiaoying Zheng,
  • Qiang Liao,
  • Yong Wang,
  • Yifan Pu,
  • Guoqiang Gao,
  • Guangning Wu

摘要

The capacity prediction of lithium battery can guide the replacement of batteries that are about to be retired, which is the most important step of battery management system (BMS). In order to predict the capacity degradation under different conditions, this paper proposes a novel method using long short-term memory (LSTM), particle swarm optimization (PSO), and singular spectrum analysis (SSA). Firstly, SSA decomposes the raw capacity degradation trend into global degraded trend, reversible degradation, and noise. Then, LSTM is adopted to learn and predict the capacity degradation. Finally, PSO automatically optimizes the initial learning rate and the number of hidden layers of LSTM for improving the prediction accuracy of battery capacity degradation. The proposed SSA-PSO-LSTM method is validated by different lithium-ion batteries from the NASA dataset. The experimental results show that both the root mean square error (RMSE) and the mean absolute error (MAE) of the prediction results are less than 1%, demonstrating the high accuracy and effectiveness of the proposed approach.