基于双向长短期记忆网络与注意力机制的锂电池剩余使用寿命预测模型

doi:10.3969/j.issn.1008-0198.2025.05.001

湖南电力 ›› 2025, Vol. 45 ›› Issue (5): 1-7.doi: 10.3969/j.issn.1008-0198.2025.05.001

• 专家专栏：配网灵活储能技术 • 下一篇

基于双向长短期记忆网络与注意力机制的锂电池剩余使用寿命预测模型

朱光明^1,2, 胡锦豪^1,3, 徐松^1,2, 万代¹

1.国网湖南省电力有限公司电力科学研究院, 湖南长沙 410208;
2.高效清洁发电技术湖南省重点实验室, 湖南长沙 410208;
3.湖南省湘电试验研究院有限公司, 湖南长沙 410208

收稿日期:2025-08-20 修回日期:2025-09-05 发布日期:2025-11-11
通信作者: 胡锦豪(1994),男,硕士,工程师,研究方向为电力电子技术。
作者简介:朱光明(1975),男,博士,教授级高级工程师,研究方向为电池储能、新能源并网、电力电子技术; 徐松(1981),男,博士,高级工程师,研究方向为电池储能技术;万代（1985）,男,高级工程师,研究方向为新能源并网、电力电子技术。
基金资助:
湖南省湘电试验研究院有限公司科技项目（5216AV24000A）

Remaining Useful Life Prediction Model Based on Bidirectional Long Short-Term Memory Network and Attention Mechanism for Lithium Batteries

ZHU Guangming^1,2, HU Jinhao^1,3, XU Song^1,2, WAN Dai¹

1. State Grid Hunan Electric Power Company Limited Research Institute, Changsha 410208, China;
2. Hunan Province Key Laboratory of Efficient and Clean Power Generation Technologies, Changsha 410208, China;
3. Hunan Xiangdian Test & Research Institute Co., Ltd., Changsha 410208, China

Received:2025-08-20 Revised:2025-09-05 Published:2025-11-11

摘要/Abstract

摘要： 为提升锂电池剩余使用寿命预测的可靠性与准确性,构建一种基于双向长短期记忆网络（bidirectional long short-term memory,BiLSTM)-注意力（Attention)机制的锂电池剩余使用寿命预测模型。该模型先以锂电池充放电过程中的关键性能参数为输入,利用BiLSTM网络的双向记忆特性,充分捕捉电池性能退化过程中的长短期时间依赖关系,有效挖掘隐藏在数据中的退化趋势特征。其次,引入Attention机制,对影响剩余使用寿命预测的关键时间步特征赋予更高权重,强化模型对重要信息的聚焦能力,进而输出精准的剩余使用寿命预测值。通过在CALCE公开的4组锂电池数据集上进行验证分析,实验结果表明,BiLSTM-Attention模型的R²值分别达到98.77%、99.64%、99.50%、98.47%,与其他预测方法相比,该方法展现出更优的预测性能。

关键词: 锂电池, 剩余使用寿命预测, BiLSTM, 注意力机制

Abstract: To enhance the reliability and accuracy of predictions, a lithium battery RUL prediction model based on Bidirectional Long Short-Term Memory Network(BiLSTM) and Attention mechanism is constructed. Firstly, the model takes the key performance parameters during the charge-discharge process of lithium batteries as the inputs, and utilizes the bidirectional memory characteristics of the BiLSTM network to fully capture the long-term and short-term temporal dependencies in the battery performance degradation process and effectively excavates the degradation trend features hidden in the data. Secondly, the Attention mechanism is introduced to assign higher weights to the key time-step features that affect RUL prediction, strengthening the model's ability to focus on important information and thereby outputting accurate RUL prediction values. Validation and analysis on four sets of public CALCE lithium battery datasets show that the R² values of the BiLSTM-Attention model reach 98.77%, 99.64%, 99.50%, and 98.47% respectively. Compared with other prediction methods, this method exhibits superior prediction performance.

Key words: lithium battery, remaining useful life (RUL) prediction, BiLSTM, attention mechanism

中图分类号:

TM911

朱光明, 胡锦豪, 徐松, 万代. 基于双向长短期记忆网络与注意力机制的锂电池剩余使用寿命预测模型[J]. 湖南电力, 2025, 45(5): 1-7.

ZHU Guangming, HU Jinhao, XU Song, WAN Dai. Remaining Useful Life Prediction Model Based on Bidirectional Long Short-Term Memory Network and Attention Mechanism for Lithium Batteries[J]. Hunan Electric Power, 2025, 45(5): 1-7.

参考文献

[1] ZHAO F M,GAO D X,CHENG Y M,et al.Application of state of health estimation and remaining useful life prediction for lithium-ion batteries based on AT-CNN-BiLSTM[J]. Scientific Reports,2024,14:29026.
[2] HU X S,FENG F,LIU K L,et al.State estimation for advanced battery management:key challenges and future trends?[J]. Renewable and Sustainable Energy Reviews,2019,114:109334.
[3] SADABADI K K,JIN X,RIZZONI G.Prediction of remaining useful life for a composite electrode lithium-ion battery cell using an electrochemical model to estimate the state of health[J]. Journal of Power Sources,2021,481:228861.
[4] YANG X Z,ZHANG H Y,LIU Q,et al.The Li-ion battery industry and its challenges[J]. Nature Reviews Chemistry,2025,9(8):497-498.
[5] WANG M,WU S M,CHEN Y,et al.Quantitative safety assessment of lithium-ion batteries:integrating abuse risks and intrinsic safety[J]. Journal of Power Sources,2025,640:236789.
[6] YAO L,XU S M,TANG A H,et al.A review of lithium-ion battery state of health estimation and prediction methods[J]. World Electric Vehicle Journal,2021,12(3):113.
[7] VIGNESH S,CHE H S,SELVARAJ J,et al.State of Health(SOH)estimation methods for second life lithium-ion battery:Review and challenges[J]. Applied Energy 2024,369:123542.
[8] CHEN L P,XIE S Q,LOPES A M,et al.A new SOH estimation method for Lithium-ion batteries based on model-data-fusion[J]. Energy,2024,286:129597.
[9] VAN C N,QUANG D T.Estimation of SOH and internal resistances of Lithium-ion battery based on LSTM network[J]. International Journal of Electrochemical Science,2023,18(6):100166.
[10] 戴海峰,王冬晨,姜波. 基于电化学阻抗谱的电池荷电状态估计[J]. 同济大学学报(自然科学版),2019,47(增刊1):95-98.
[11] ZHANG Y W,TANG Q C,ZHANG Y,et al.Identifying degradation patterns of lithium-ion batteries from impedance spectroscopy using machine learning[J]. Nature Communications,2020,11:1706.
[12] 王帅,韩伟,陈黎飞,等. 基于粒子滤波的锂离子电池剩余寿命预测[J]. 电源技术,2020,44(3):346-351.
[13] 吴晓丹,范波,王建祥,等. 基于VMD-TCN-Attention的锂电池寿命预测[J]. 电源技术,2023,47(10):1319-1325.
[14] LU J H,XIONG R,TIAN J P,et al.Battery degradation prediction against uncertain future conditions with recurrent neural network enabled deep learning[J]. Energy Storage Materials,2022,50:139-151.
[15] UNGUREAN L,MICEA M V,CARSTOIU G.Online state of health prediction method for lithium-ion batteries based on gated recurrent unit neural networks[J]. International Journal of Energy Research,2020,44(8):6767-6777.
[16] QU J T,LIU F,MA Y X,et al.A neural-network- based method for RUL prediction and SOH monitoring of lithium-ion battery[J]. IEEE Access,2019,7:87178-87191.
[17] LI P,ZHANG Z J,XIONG Q Y,et al.State-of-health estimation and remaining useful life prediction for the lithium-ion battery based on a variant long short term memory neural network[J]. Journal of Power Sources, 2020,459:228069.
[18] WANG F K,HUANG C Y,MAMO T.Ensemble model based on stacked long short-term memory model for cycle life prediction of lithium-ion batteries[J]. Applied Sciences,2020,10(10):3549.
[19] HOCHREITER S,SCHMIDHUBER J.Long short-term memory[J]. Neural Computation,1997,9(8):1735-1780.
[20] SCHUSTER M,PALIWAL K K.Bidirectional recurrent neural networks[J]. IEEE transactions on Signal Processing,1997,45(11):2673-2681.
[21] HUANG Z Z,LIANG S W,LIANG M F.A generic shared attention mechanism for various backbone neural networks[J]. Neurocomputing,2025,611:128697.
[22] TAO S Y,MA R F,ZHAO Z X,et al.Generative learning assisted state-of-health estimation for sustainable battery recycling with random retirement conditions[J]. Nature Communications,2024,15:10154.

基于双向长短期记忆网络与注意力机制的锂电池剩余使用寿命预测模型

Remaining Useful Life Prediction Model Based on Bidirectional Long Short-Term Memory Network and Attention Mechanism for Lithium Batteries

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 6

编辑推荐

Metrics

本文评价

[1]	刘晶菊, 陈罗甲, 王江峰, 陈宝辉, 吴传平. 隔膜胶层材料对磷酸铁锂电池安全性的影响[J]. 湖南电力, 2025, 45(5): 41-46.
[2]	王生祺, 陈海波, 叶金翔. 基于深度学习与局部感知注意力的避雷器套管红外图像分割方法[J]. 湖南电力, 2025, 45(2): 122-128.
[3]	于惠钧, 夏梦, 陈刚, 谭福元, 徐银凤. 基于牛顿-拉弗森优化算法与注意力机制优化TCN-GRU的短期电力负荷预测[J]. 湖南电力, 2024, 44(6): 120-127.
[4]	徐松, 万涛, 李欣, 吴俊杰, 查方林, 魏加强, 蔡宇峰, 刘奕奕. 基于声学检测技术的刀片储能锂电池热失控行为研究[J]. 湖南电力, 2024, 44(5): 17-23.
[5]	赵文钦, 肖辉, 张茂林, 李逸航, 龙飞宇, 文章. 变电站锂电池并联系统与级差配合研究[J]. 湖南电力, 2022, 42(4): 23-27.
[6]	苏婷婷, 彭贺翔, 王灿, 李波, 廖凯, 刘世峰. 基于BiLSTM的电压暂降原因辨识方法研究[J]. 湖南电力, 2022, 42(1): 32-37.