| 45 | 0 | 14 |
| 下载次数 | 被引频次 | 阅读次数 |
准确预测电池健康状态（state of health, SOH）是保障电动汽车电池系统安全可靠运行的关键。现有方法在不同电池化学体系下预测SOH的通用性仍然不足，针对该问题，本研究以镍钴铝、镍钴锰和磷酸铁锂三种典型锂离子电池为研究对象，提出一种新型的并行加权预测架构，将科尔莫戈洛夫—阿诺尔德网络（kolmogorov-arnold network, KAN）与物理信息神经网络（physics-informed neural networks, PINN）集成构建为K-PINN模型。该模型仅需单一电压信号作为输入特征，即可直接适配不同化学体系的电池的SOH预测，显著降低了对多源数据的依赖。实验结果表明，K-PINN框架展现出超强的泛化性能。以镍钴锰电池为例，该模型的平均绝对误差达到0.009，相较于现有数据驱动技术，其预测精度实现了显著提升。所以，K-PINN可以为电池健康状态监测提供新的解决方案，为推动电池管理系统优化提供了技术支撑。
Abstract：Accurate prediction of the state of health(SOH) of batteries is crucial for ensuring the safe and reliable operation of electric vehicle battery systems. However, existing methods still lack generalizability in predicting SOH across different battery chemistries. To address this issue, this study proposes a novel parallel weighted prediction architecture that integrates the Kolmogorov-Arnold network(KAN) with physics-informed neural networks(PINN) to construct a K-PINN model. This model only requires a single voltage signal as the input feature and can be directly adapted for SOH prediction of batteries with different chemical systems, significantly reducing reliance on multi-source data. Experimental results demonstrate that the K-PINN framework exhibits superior generalization performance. Its prediction accuracy is significantly higher than that of existing data-driven techniques, particularly on the nickel-cobalt-manganese battery dataset, where the K-PINN achieves a mean absolute error of only 0.009. In conclusion, this study verifies that K-PINN model can serve as an effective solution for battery SOH monitoring, providing valuable technical support for the optimization of battery management systems.
[1]WANG L, RAN X, XU S,et al.ADSTAN:Adversarial dynamic spatiotemporal attention networks for unsupervised cross-domain battery SOH estimation[J].IEEE Transactions on Industrial Informatics,2025,21(10):7575-86.
[2]QIN Y, YUEN C, YIN X,et al.A transferable multi-stage model with cycling discrepancy learning for lithium-ion battery state of health estimation[J].IEEE Transactions on Industrial Informatics,2023,19(2):1933-46.
[3]王洋,邢丽坤.基于蜜獾优化算法RVM电池健康状态估计[J/OL].重庆工商大学学报(自然科学版),1-12[2025-11-12].https://link.cnki.net/urlid/50.1155.N.20250916.1749.005.
[4]LIN C, WU L, TUO X,et al.A lightweight two-stage physics-informed neural network for SOH estimation of lithium-ion batteries with different chemistries[J].Journal of Energy Chemistry, 2025,105(6):261-279.
[5]FENG Y, XUE C, HAN F, et al. State-of-charge and state-of-health estimation in Li-ion batteries using cascade electrochemical model-based sliding-mode observers[J]. Batteries,2024,10(8):290.
[6]任恒起.基于分数阶模型的多时间尺度锂电池状态在线估计研究[D].淮南:安徽理工大学,2025.
[7]LUO P, LI P, MA D, et al. A novel capacity fade model of lithium-ion cells considering the influence of stress[J].Journal of the Electrochemical Society,2021,168(9):090537.
[8]LI J, ADEWUYI K, LOTFI N, et al. A single particle model with chemical/mechanical degradation physics for lithium ion battery state of health(SOH)estimation[J].Applied Energy,2018,212:1178-1190.
[9]黄亮亮,郭拯诗,何峰,等.基于WPD-GA-BP的电动汽车动力电池健康状态预测方法[J].华东交通大学学报,2026,43(1):101-113.
[10]DUNG L R , WU S H , YUAN H F. An SOH estimation system based on time-constant-ratio measuremen[J].IEEE23rd International Symposium on Industrial Electronics(ISIE),2014:1784-1787.
[11]韦荣阳.基于CC-CV充电模式的锂离子电池健康状态估计方法研究[D].杭州:浙江大学,2023.
[12]LUO K, ZHENG H, SHI Z. A simple feature extraction method for estimating the whole life cycle state of health of lithium-ion batteries using transformer-based neural network[J]. Journal of Power Sources,2023,576:233139.
[13]WU X, LI P, DENG Z, et al. LDNet-RUL:Lightweight deformable neural network for remaining useful life prognostics of lithium-ion batteries[J]. IEEE Transactions on Power Electronics,2025,40(9):13514-28.
[14]WU X, LI P, ZHANG Y, et al. Knowledge distillation-based lightweight deformable network for remaining useful life prognostics of vehicle power battery[J]. Energy,2025,337:138522.
[15]陈嘉铭,唐文俊,何山,等.基于短时充电片段的实车动力电池SOH估计[J].广西科技大学学报,2025,36(4):67-73.
[16]张慧颖,王军华,丁汀,等.基于多指标的CNN-LSTM锂电池SOH估计模型[J].电源技术,2023,47(10):1303-1307.
[17]严芷涵,王学远,魏学哲,等.基于电化学阻抗谱几何解析的锂离子电池健康状态评估[J].储能科学与技术,2025,14(12):4732-4742.
[18]沈建明,孙飞,汤碧君,等.多重机理约束的锂电池对抗解耦表征与健康估计方法[J].控制工程,1-7[2025-11-12].https://doi.org/10.14107/j.cnki.kzgc.20250290.
[19]WEN P, YE Z S, LI Y, et al. Physics-informed neural networks for prognostics and health management of lithium-ion batteries[J]. IEEE Transactions on Intelligent Vehicles,2023,9(1):2276-2289.
[20]潘大为,师杰,杜宇航,等.基于BNN-PF的卫星锂离子电池多工况SOH估计[J].电子测量与仪器学报,2024,38(9):104-115.
[21]胡钰航,廖宇,崔琨,等.基于SA-CDC-GRU-AE模型的锂离子电池健康状态估计方法[J].湖北民族大学学报(自然科学版),2025,43(2):266-271.
[22]董静,金帅.基于ICA与GS-SVM的电池健康状态估计[J].重庆理工大学学报(自然科学),2025.39(4):17-26
[23]ZHU J, WANG Y, HUANG Y, GOPALUNI R B,et al.Data-driven capacity estimation of commercial lithium-ion batteries from voltage relaxation[J].Nature Communications,2022,13(1):2261.
[24]SEVERSON K A , ATTIA P M , JIN N ,et al.Data-driven prediction of battery cycle life before capacity degradation[J].Nature Energy,2019,4(5):383-91.
[25]HUANG K, GUO Y F, LI S M. Estimation of maximum available capacity of lithium-ion battery based on multi-view features extracted from reconstructed charging curve[J]. International Journal of Hydrogen Energy,2022,47(44):19175-19194.
基本信息:
中图分类号:TP183;TM912
引用信息:
[1]周潮湧,汪紫薇,高阳,等.基于KAN与物理信息神经网络融合的动力电池健康状态预测研究[J].井冈山大学学报(自然科学版),2026,47(03):87-96.
基金信息:
国家自然科学基金项目(52275146,61804054,12411530109,12174102)
2026-05-10
2026-05-10