基于自适应分频混合网络的锂电池SOH预测方法

马正; 刘景霞; 杨培宏; 张宝华

doi:10.19799/j.cnki.2095-4239.2026.0713

您当前的位置：

首页 >

文章列表页 >

基于自适应分频混合网络的锂电池SOH预测方法

XXXX | 更新时间：2026-10-13

- 基于自适应分频混合网络的锂电池SOH预测方法
- Lithium-Ion Battery SOH Prediction Based on an Adaptive Frequency-Division Hybrid Network
- 储能科学与技术 2026年页码：1-15
- 作者机构：
  
  内蒙古科技大学自动化与电气工程学院，内蒙古包头 014010
- 作者简介：
  
  马正（2002—），男，硕士研究生，研究方向为锂电池健康状态估计，E-mail：MZheng1763@163.com；
  刘景霞，副教授，研究方向为储能电池技术、风功率预测等，E-mail：958428915@qq.com。
- 基金信息：
  
  内蒙古自治区自然科学基金项目(2025MS05086);内蒙古自治区重点研发与成果转化项目(2026YFDZ0150);内蒙古自治区重大专项(2026ZD0132)
- DOI：10.19799/j.cnki.2095-4239.2026.0713
  中图分类号： TM 912
- 收稿：2026-08-14，
  
  修回：2026-10-04，
  
  网络首发：2026-10-13，
- 稿件说明：
移动端阅览
马正, 刘景霞, 杨培宏, 等. 基于自适应分频混合网络的锂电池SOH预测方法[J]. 储能科学与技术, XXXX, XX(XX): 1-15.

MA Zheng, LIU Jingxia, YANG Peihong, et al. Lithium-Ion Battery SOH Prediction Based on an Adaptive Frequency-Division Hybrid Network[J]. Energy Storage Science and Technology, XXXX, XX(XX): 1-15.
马正, 刘景霞, 杨培宏, 等. 基于自适应分频混合网络的锂电池SOH预测方法[J]. 储能科学与技术, XXXX, XX(XX): 1-15. DOI： 10.19799/j.cnki.2095-4239.2026.0713.

MA Zheng, LIU Jingxia, YANG Peihong, et al. Lithium-Ion Battery SOH Prediction Based on an Adaptive Frequency-Division Hybrid Network[J]. Energy Storage Science and Technology, XXXX, XX(XX): 1-15. DOI： 10.19799/j.cnki.2095-4239.2026.0713.

摘要

精确预测锂电池健康状态(state of health，SOH)对于保障电池安全运行、制定维护策略及开展寿命管理具有重要意义。针对原始SOH序列具有非线性、非平稳和多时间尺度特征，以及不同频率退化分量采用统一模型难以实现针对性表征的问题，提出一种基于充电阶段数据和自适应分频混合网络的锂电池跨电池SOH预测方法。首先，仅利用训练电池充电阶段数据提取能够表征电池退化状态的健康特征(health indicators，HIs)，并采用变分模态分解(variational mode decomposition，VMD)对训练电池SOH序列进行多尺度分解；进一步利用排列熵(permutation entropy，PE)量化不同本征模态分量的复杂度，并结合K-means聚类实现高、低频分量的自适应划分。针对不同复杂度分量的动态特性，采用灰叶猴寻优算法(gray langurs optimizer，GLO)优化的双向长短期记忆网络-注意力机制(BiLSTM-Attention)预测高频分量，采用结构相对简洁的门控循环单元(gated recurrent unit，GRU)预测低频分量。模型训练过程中，VMD分解、特征筛选、参数寻优及模型训练均仅基于训练电池完成；跨电池测试阶段固定全部模型参数，仅利用测试电池截至当前循环已经获得的历史HIs，一步预测下一循环对应的各模态分量，并通过重构得到SOH预测结果。以CALCE-CS2系列电池为对象，首先采用CS2_35作为唯一训练电池，CS2_36、CS2_37和CS2_38作为独立测试电池开展基准实验，并通过依次更换唯一训练电池进行交叉验证。实验结果表明，所提方法能够有效表征SOH退化过程中的多时间尺度信息，在不同训练电池条件下均保持较为稳定的预测性能，进一步验证了其在CALCE-CS2同系列电池间的跨电池预测能力。

Abstract

Accurate prediction of lithium-ion battery state of health (SOH) provides an important basis for safe operation

maintenance decision-making

and lifetime management. To address the nonlinear

nonstationary

and multi-timescale characteristics of battery SOH degradation

as well as the insufficient representation caused by applying a unified prediction model to components with different dynamic complexities

an adaptive frequency-division hybrid neural network is proposed. First

health indicators (HIs) that characterize battery degradation are extracted from the charging-stage data of the training battery. Variational mode decomposition (VMD) is then applied only to the SOH sequence of the training battery to separate degradation information at multiple time scales. Permutation entropy (PE) is employed to quantify the complexity of different intrinsic mode functions (IMFs)

and K-means clustering is further used to adaptively divide them into high- and low-frequency components. According to their different dynamic characteristics

a bidirectional long short-term memory-attention (BiLSTM-Attention) model optimized by the Gray Langurs Optimizer (GLO) is used for high-frequency components

while a gated recurrent unit (GRU) model is used for low-frequency components. During model development

HI selection

VMD decomposition

hyperparameter optimization

and model training are performed exclusively using the training battery. During cross-cell testing

all model parameters are fixed

and only the historical HIs available up to the current cycle of the test battery are used to predict the intrinsic mode functions of the next cycle in a one-step-ahead manner

followed by reconstruction of the final SOH. Experiments are conducted on the CALCE-CS2 battery series. CS2_35 is first used as the sole training cell and CS2_36

CS2_37

and CS2_38 are used as independent test cells

followed by cross-validation by alternately selecting each cell as the sole training cell. The results show that the proposed method can effectively characterize the multi-timescale information in the SOH degradation process and maintain relatively stable prediction performance under different training-cell conditions

further validating its cross-cell prediction capability within the CALCE-CS2 series.

关键词

Keywords

references

来鑫,陈权威,顾黄辉,等. 面向"双碳"战略目标的锂离子电池生命周期评价:框架、方法与进展[J]. 机械工程学报, 2022, 58(22): 3-18.

LAI X, CHEN Q W, GU H H, et al. Life cycle assessment of lithium-ion batteries toward the "dual-carbon" strategic goals: Framework, methods and progress[J]. Journal of Mechanical Engineering, 2022, 58(22): 3-18.

蒋帅,陈铖,段砚州,等. 温度和老化影响下的锂离子电池荷电状态和健康状态联合估计方法研究[J]. 机械工程学报, 2024, 60(18): 266-275.

JIANG S, CHEN C, DUAN Y Z, et al. Joint estimation of state of charge and state of health of lithium-ion batteries considering temperature and aging effects[J]. Journal of Mechanical Engineering, 2024, 60(18): 266-275.

王怡,陈学兵,王愿习,等. 储能锂离子电池多层级失效机理及分析技术综述[J]. 储能科学与技术, 2023, 12(7): 2079-2094.

WANG Y, CHEN X B, WANG Y X, et al. Review of multilevel failure mechanisms and analysis technologies for energy storage lithium-ion batteries[J]. Energy Storage Science and Technology, 2023, 12(7): 2079-2094.

李放,闵永军,张涌. 基于大数据的动力锂电池可靠性关键技术研究综述[J]. 储能科学与技术, 2023, 12(6): 1981-1994.

LI F, MIN Y J, ZHANG Y. Review of key reliability technologies for power lithium-ion batteries based on big data[J]. Energy Storage Science and Technology, 2023, 12(6): 1981-1994.

熊庆,邸振国,汲胜昌. 锂离子电池健康状态估计及寿命预测研究进展综述[J]. 高电压技术, 2024, 50(3): 1182-1195.

XIONG Q, DI Z G, JI S C. Review of research progress in state-of-health estimation and lifetime prediction of lithium-ion batteries[J]. High Voltage Engineering, 2024, 50(3): 1182-1195.

金帅,董静.基于数据驱动的锂离子电池健康状态估计研究进展综述[J]. 仪器仪表学报, 2024, 45(3): 45-59.

JIN S, DONG J. Review of research progress in data-driven state-of-health estimation of lithium-ion batteries[J]. Chinese Journal of Scientific Instrument, 2024, 45(3): 45-59.

黎冲,王成辉,王高,等. 基于数据驱动的锂离子电池健康状态估计技术[J]. 中国电力, 2022, 55(8): 73-86, 95.

LI C, WANG C H, WANG G, et al. Data-driven state-of-health estimation technology for lithium-ion batteries[J]. Electric Power, 2022, 55(8): 73-86, 95.

朱振宇,高德欣. 基于CNN-BiLSTM网络的锂离子电池健康状态检测方法[J]. 电子测量技术, 2023, 46(3): 128-133.

ZHU Z Y, GAO D X. State-of-health detection method for lithium-ion batteries based on CNN-BiLSTM network[J]. Electronic Measurement Technology, 2023, 46(3): 128-133.

Guo Y, Yang D, Zhao K, et al. State of health estimation for lithium-ion battery based on Bi-directional long short-term memory neural network and attention mechanism[J]. Energy Reports, 2022, 8: 208–215.

Liu K, Kang L, Xie D. Online State of Health Estimation of Lithium-Ion Batteries Based on Charging Process and Long Short-Term Memory Recurrent Neural Network[J]. Batteries, 2023, 9(2): 94.

Li Y, Luo L, Zhang C, et al. State of Health Assessment for Lithium-Ion Batteries Using Incremental Energy Analysis and Bidirectional Long Short-Term Memory[J]. World Electric Vehicle Journal, 2023, 14(7): 188.

吴铁洲,朱俊超,成雄帆,等. 基于充电阶段数据与GWO-BiLSTM模型的锂电池SOH估计方法[J]. 电源技术, 2024, 48(11): 2184-2194.

WU T Z, ZHU J C, CHENG X F, et al. State-of-health estimation method for lithium-ion batteries based on charging-stage data and GWO-BiLSTM model[J]. Chinese Journal of Power Sources, 2024, 48(11): 2184-2194.

Ma Y, Li J, Gao J, et al. State of health prediction of lithium-ion batteries under early partial data based on IWOA-BiLSTM with single feature[J]. Energy, 2024, 295: 131085.

何山,赵宇明,田爱娜,等. 基于充电片段和PSO-BP的锂电池SOH在线估计方法[J]. 电源技术, 2025, 49(2): 383-389.

HE S, ZHAO Y M, TIAN A N, et al. Online state-of-health estimation method for lithium-ion batteries based on charging segments and PSO-BP[J]. Chinese Journal of Power Sources, 2025, 49(2): 383-389.

高国章,刘长昊,袁裕鹏,等. 基于PKO-XGBoost-TCN的锂电池SOH估计[J]. 电源技术, 2025, 49(12): 2553-2562.

GAO G Z, LIU C H, YUAN Y P, et al. State-of-health estimation of lithium-ion batteries based on PKO-XGBoost-TCN[J]. Chinese Journal of Power Sources, 2025, 49(12): 2553-2562.

Shabrawy M, El-Kenawy E.-S. M, Abdel-Hamid N. B, et al. An optimization-driven hierarchical deep learning approach using the Gray Langurs algorithm for data-driven seismic activity prediction[J]. Scientific Reports, 2026, 16: 18846.

XIA F, WANG K, CHEN J. State-of-Health Prediction for Lithium-Ion Batteries Based on Complete Ensemble Empirical Mode Decomposition with Adaptive Noise-Gate Recurrent Unit Fusion Model[J]. Energy Technology, 2022, 10(4): 2100767.

FU J, WU C, WANG J, et al. Lithium-ion Battery SOH Prediction Based on VMD-PE and Improved DBO Optimized Temporal Convolutional Network Model[J]. Journal of Energy Storage, 2024, 87: 111392.

HUI Z, SHI Z, WANG R, et al. Health Prediction of Lithium-ion Batteries by Combining with Empirical Mode Decomposition and PF-GPR Algorithm[J]. Materials Today Energy, 2024, 42: 101562.

Su Y X, Wang Z Y, Dong Z C, et al. Frequency-aware ultra-short-term wind power forecasting using CEEMDAN–VMD–SE and Transformer–GRU networks[J]. Energy, 2025, 338: 138715.

杨朋朋,曾圣浩,薛海,等. 基于双策略优化VMD-HO-LSTM的锂电池剩余寿命预测[J]. 重庆交通大学学报(自然科学版), 2026, 45(1): 120-128.

YANG P P, ZENG S H, XUE H, et al. Remaining useful life prediction of lithium-ion batteries based on dual-strategy optimized VMD-HO-LSTM[J]. Journal of Chongqing Jiaotong University (Natural Science), 2026, 45(1): 120-128.

田成文,孙丙香,赵鑫泽,等. 基于数据驱动的锂离子电池快速寿命预测[J]. 储能科学与技术, 2024, 13(9): 3103-3111.

TIAN C W, SUN B X, ZHAO X Z, et al. Data-driven rapid lifetime prediction of lithium-ion batteries[J]. Energy Storage Science and Technology, 2024, 13(9): 3103-3111.

ZHANG L, JI T, YU S, et al. Accurate prediction approach of SOH for lithium-ion batteries based on LSTM method[J]. Batteries, 2023, 9(3): 177.

LI Z, ZHANG X, GAO W. State of health estimation of lithium-ion battery during fast charging process based on BiLSTM-Transformer[J ] . Energ y , 2024, 311: 133418.

Barshandeh S, Khodadadi N, Abdollahzadeh B, et al. Gray langurs optimizer: a multi-group bio-inspired optimization algorithm[J]. Artificial Intelligence Review, 2026, 59: 166.

刘芊彤,邢远秀. 基于VMD-PSO-GRU模型的锂离子电池剩余寿命预测[J]. 储能科学与技术, 2023, 12(1): 236-246.

LIU Q T, XING Y X. Remaining useful life prediction of lithium-ion batteries based on VMD-PSO-GRU model[J]. Energy Storage Science and Technology, 2023, 12(1): 236-246.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于白鹭群优化高斯过程回归的锂电池SOH估计方法

基于自适应最优组合核函数高斯过程回归的锂电池健康状态区间估计

一种基于不一致性特征融合的电池组健康状态估计方法

粒径级配对干法电极阴极材料结构和性能的影响

基于目标鲁棒的基站储能联合参与电力市场的数学决策模型