Scholarlink

Accurately predicting the remaining useful life (RUL) of lithium-ion batteries is of vital significance for guaranteeing the safety and reliability of power systems. In view of the nonlinear characteristics of the aging trajectory of lithium-ion battery, a deep learning model combining the Recursive Ensemble Empirical Mode Decomposition (REMD) algorithm and residual Gated Recurrent Unit (resGRU) for predicting lithium-ion battery RUL is proposed in this paper. Firstly, the REMD algorithm is employed to decompose the aging data of lithium-ion batteries, and the Intrinsic Mode Functions (IMFs) and residual sequences are extracted. Then, the residual sequences are predicted based on a Dual-layer CNN-resGRU prediction model. IMFs are predicted based on Gaussian Process Regression (GPR). Finally, the aging trajectory of lithium-ion battery is calculated by fusing the prediction results from IMFs and residual sequences. Experimental results demonstrated that the RUL prediction method proposed in this research can precisely track the aging trajectory of lithium-ion batteries.

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

Lithium-Ion Battery RUL Prediction Method Based on REMD and Dual-Layer CNN-resGRU

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Guoqing Hua,
  • Yupeng Wu,
  • Min Xie,
  • Yang Chen,
  • Hongliang Sima,
  • Mengling Liu,
  • Yufeng Zhang,
  • Chaolong Zhang

摘要

Accurately predicting the remaining useful life (RUL) of lithium-ion batteries is of vital significance for guaranteeing the safety and reliability of power systems. In view of the nonlinear characteristics of the aging trajectory of lithium-ion battery, a deep learning model combining the Recursive Ensemble Empirical Mode Decomposition (REMD) algorithm and residual Gated Recurrent Unit (resGRU) for predicting lithium-ion battery RUL is proposed in this paper. Firstly, the REMD algorithm is employed to decompose the aging data of lithium-ion batteries, and the Intrinsic Mode Functions (IMFs) and residual sequences are extracted. Then, the residual sequences are predicted based on a Dual-layer CNN-resGRU prediction model. IMFs are predicted based on Gaussian Process Regression (GPR). Finally, the aging trajectory of lithium-ion battery is calculated by fusing the prediction results from IMFs and residual sequences. Experimental results demonstrated that the RUL prediction method proposed in this research can precisely track the aging trajectory of lithium-ion batteries.