基于TWP-SVR的锂离子电池健康状态估计

doi:10.19799/j.cnki.2095-4239.2022.0184

摘要/Abstract

摘要：

健康状态(state of health，SOH)是评估锂离子电池老化程度和剩余使用寿命的重要指标。然而，SOH无法通过直接测量获得，本工作提出了一种基于时间规整图(time warp profile，TWP)提取间接健康特征参数，使用支持向量机回归(vector machine regression，SVR)模型估计SOH的方法。首先，通过TWP将锂离子电池不同循环充放电压曲线转换为相位差异曲线。然后，从相位差异曲线中提取出4个间接健康特征。接着，采用线性核函数的SVR模型估计SOH。最后，以美国航空航天局(National Aeronautics and Space Administration，NASA)、美国保险商实验室公司和普渡大学(Underwriters Laboratories Inc.-Purdue University，UL-PUR)的开源数据集和储能电站实测数据进行验证。其中，储能电站数据实验结果表明，TWP-SVR模型的均方根误差(root mean square error，RMSE)、平均绝对误差(mean absolute error，MAE)的样本标准差(sample standard deviation，SSD)小于0.0015，四分位距(inter-quartile range，IQR)小于0.0022；平均绝对百分比误差(mean absolute percentage error，MAPE)的SSD和IQR分别为0.0152和0.0220，表明所提TWP-SVR方法保持较高准确性的同时具有良好的稳定性。

关键词: 锂离子电池, 健康状态, 时间规整图, 支持向量机回归

Abstract:

State of Health (SOH) is an important index for evaluating the aging degree and remaining service life of lithium-ion batteries. However, SOH cannot be obtained by direct measurement. This study proposes a method to extract indirect health feature parameters based on time warp profile (TWP) and estimate the SOH using a vector machine regression (SVR) model. First, TWP converts the different cycle charge-discharge voltage curves of a lithium-ion battery into a phase difference curve. Then, four indirect health features are extracted from the phase difference curve. Further, SOH is estimated using the SVR model of the linear kernel function. Finally, it is verified by measured data of energy storage power station and the open-source data sets of NASA, Underwriters Laboratories Inc.-Purdue University (UL-PUR). The experimental results of energy storage power station data show that the sample standard deviation (SSD) of the root mean square error and mean absolute error of the TWP-SVR model is less than 0.0015, and the interquartile range (IQR) is less than 0.0022; the SSD and IQR of mean absolute percentage error are 0.0152 and 0.0220, respectively, indicating that the proposed TWP-SVR method maintains high accuracy and good stability.

Key words: lithium ion battery, status of health, time warp profile, support vector machine regression

中图分类号:

TM 911

韦荣阳, 毛阗, 高晗, 彭建仁, 杨健. 基于TWP-SVR的锂离子电池健康状态估计[J]. 储能科学与技术, 2022, 11(8): 2585-2599.

Rongyang WEI, Tian MAO, Han GAO, Jianren PENG, Jian YANG. Health state estimation of lithium ion battery based on TWP-SVR[J]. Energy Storage Science and Technology, 2022, 11(8): 2585-2599.

图/表 24

图 1

图 2

图 3

表1

图 4

表 2

表 3

图 5

图 6

图 7

图 8

图 9

图 10

图 11

图 12

图 13

图 14

图 15

表 4

表 5

图 16

图 17

图 18

表 6

参考文献 27

1	杨杰, 王婷, 杜春雨, 等. 锂离子电池模型研究综述[J]. 储能科学与技术, 2019, 8(1): 58-64.
	YANG J, WANG T, DU C Y, et al. Overview of the modeling of lithium-ion batteries[J]. Energy Storage Science and Technology, 2019, 8(1): 58-64.
2	苏伟, 钟国彬, 沈佳妮, 等. 锂离子电池故障诊断技术进展[J]. 储能科学与技术, 2019, 8(2): 225-236.
	SU W, ZHONG G B, SHEN J N, et al. The progress in fault diagnosis techniques for lithium-ion batteries[J]. Energy Storage Science and Technology, 2019, 8(2): 225-236.
3	卢婷, 杨文强. 锂离子电池全生命周期内评估参数及评估方法综述[J]. 储能科学与技术, 2020, 9(3): 657-669.
	LU T, YANG W Q. Review of evaluation parameters and methods of lithium batteries throughout its life cycle[J]. Energy Storage Science and Technology, 2020, 9(3): 657-669.
4	徐超, 李立伟, 杨玉新, 等. 基于改进粒子滤波的锂电池SOH预测[J]. 储能科学与技术, 2020, 9(6): 1954-1960.
	XU C, LI L W, YANG Y X, et al. Lithium-ion battery SOH estimation based on improved particle filter[J]. Energy Storage Science and Technology, 2020, 9(6): 1954-1960.
5	ZHENG L F, ZHANG L, ZHU J G, et al. Co-estimation of state-of-charge, capacity and resistance for lithium-ion batteries based on a high-fidelity electrochemical model[J]. Applied Energy, 2016, 180: 424-434.
6	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.
7	WENG C H, CUI Y J, SUN J, et al. On-board state of health monitoring of lithium-ion batteries using incremental capacity analysis with support vector regression[J]. Journal of Power Sources, 2013, 235: 36-44.
8	BERECIBAR M, GARMENDIA M, GANDIAGA I, et al. State of health estimation algorithm of LiFePO₄ battery packs based on differential voltage curves for battery management system application[J]. Energy, 2016, 103: 784-796.
9	MERLA Y, WU B, YUFIT V, et al. Novel application of differential thermal voltammetry as an in-depth state-of-health diagnosis method for lithium-ion batteries[J]. Journal of Power Sources, 2016, 307: 308-319.
10	HAIDER S N, ZHAO Q C, LI X L. Cluster-based prediction for batteries in data centers[J]. Energies, 2020, 13(5): 1085.
11	ZHANG Q C, LI X Z, DU Z C, et al. Aging performance characterization and state-of-health assessment of retired lithium-ion battery modules[J]. Journal of Energy Storage, 2021, 40: doi.org/10.1016/j.est.2021.102743.
12	ZHOU K Q, QIN Y, LAU B P L, et al. Lithium-ion battery state of health estimation based on cycle synchronization using dynamic time warping[C]//IECON 2021-47th Annual Conference of the IEEE Industrial Electronics Society. October 13-16, 2021, Toronto, ON, Canada. IEEE, 2021: 1-6.
13	李海林, 梁叶, 王少春. 时间序列数据挖掘中的动态时间弯曲研究综述[J]. 控制与决策, 2018, 33(8): 1345-1353.
	LI H L, LIANG Y, WANG S C. Review on dynamic time warping in time series data mining[J]. Control and Decision, 2018, 33(8): 1345-1353.
14	FOLGADO D, BARANDAS M, MATIAS R, et al. Time alignment measurement for time series[J]. Pattern Recognition, 2018, 81: 268-279.
15	SIOROS G, NYMOEN K. Accurate shape and phase averaging of time series through Dynamic Time Warping[EB/OL]. https://arxiv.org/abs/2109.00978.
16	CHEN Z, SONG X, XIAO R, et al. State of health estimation for lithium-ion battery based on long short term memory networks[J]. DEStech Transactions on Environment Energy, 2019: doi: 10.12783/dteees/iceee2018/27855.
17	张少凤, 张清勇, 杨叶森, 等. 基于滑动窗口和LSTM神经网络的锂离子电池建模方法[J]. 储能科学与技术, 2022, 11(1): 228-239.
	ZHANG S F, ZHANG Q Y, YANG Y S, et al. Lithium-ion battery model based on sliding window and long short term memory neural network[J]. Energy Storage Science and Technology, 2022, 11(1): 228-239.
18	UNGUREAN L, MICEA M V, CÂRSTOIU 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): doi:10.1002/er.5413.
19	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.
20	王凡, 史永胜, 刘博亲, 等. 基于注意力改进BiGRU的锂离子电池健康状态估计[J]. 储能科学与技术, 2021, 10(6): 2326-2333.
	WANG F, SHI Y S, LIU B Q, et al. Health state estimation of lithium-ion batteries based on attention augmented BiGRU[J]. Energy Storage Science and Technology, 2021, 10(6): 2326-2333.
21	FAN Y X, XIAO F, LI C R, et al. A novel deep learning framework for state of health estimation of lithium-ion battery[J]. Journal of Energy Storage, 2020, 32: doi: 10.1016/j.est.2020.101741.
22	戴彦文, 于艾清. 基于健康特征参数的CNN-LSTM&GRU组合锂电池SOH估计[J]. 储能科学与技术, 2022, 11(5): 1641-1649.
	DAI Y W, YU A Q. Combined CNN-LSTM and GRU based health feature parameters for lithium-ion batteries SOH estimation[J]. Energy Storage Science and Technology, 2022, 11(5): 1641-1649.
23	FENG H L, SHI G L. SOH and RUL prediction of Li-ion batteries based on improved Gaussian process regression[J]. Journal of Power Electronics, 2021, 21(12): 1845-1854.
24	韩云飞, 谢佳, 蔡涛, 等. 结合高斯过程回归与特征选择的锂离子电池容量估计方法[J]. 储能科学与技术, 2021, 10(4): 1432-1438.
	HAN Y F, XIE J, CAI T, et al. Capacity estimation of lithium-ion batteries based on Gaussian process regression and feature selection[J]. Energy Storage Science and Technology, 2021, 10(4): 1432-1438.
25	LI R, LI W R, ZHANG H N, et al. On-line estimation method of lithium-ion battery health status based on PSO-SVM[J]. Frontiers in Energy Research, 2021, 9: doi: 10.3389/fenrg.2021.693249.
26	B S, K G. Prognostics center of excellence-data repository[EB/OL]. https://ti.arc.nasa.gov/tech/dash/groups/pcoe/prognostic-data-repository/#battery, 2007/2022-01-29.
27	JUAREZ-ROBLES D, JEEVARAJAN J A, MUKHERJEE P P. Degradation-safety analytics in lithium-ion cells: Part I. aging under charge/discharge cycling[J]. Journal of the Electrochemical Society, 2020, 167(16): doi: 10.1149/1945-7111/abc8c0.

来源	正极材料	规格	额定容量/mAh	电池数目	实验温度/℃	循环次数范围
NASA^[26]	—	18650	2000	4	24	130～170
UL-PUR^[27]	NCA	18650	3400	21	23	170～800
储能电站	LFP	模组	105000	216	21～30	160

电池	相关系数	间接健康特征
电池	相关系数	F¹	F²	F³	F⁴
5	Pearson	0.9949	0.9978	0.9926	0.9949
	Spearman	0.9725	0.9926	0.9701	0.9826
	Kendall	0.9013	0.9520	0.8959	0.9359
6	Pearson	0.9933	0.9988	0.9948	0.9942
	Spearman	0.9989	0.9990	0.9985	0.9993
	Kendall	0.9805	0.9825	0.9773	0.9859
7	Pearson	0.9891	0.9969	0.9874	0.9872
	Spearman	0.9642	0.9897	0.9637	0.9767
	Kendall	0.8811	0.9397	0.8786	0.9179
18	Pearson	0.9895	0.9972	9.8550	0.9908
	Spearman	0.9951	0.9952	0.9938	0.9971
	Kendall	0.9537	0.9576	0.9490	0.9686

算法	输入数据	特征提取算法	输出特征	拟合模型参数
TWP-SVR	充放电压曲线	TWP	F¹, F², F³, F⁴	1. 核函数：线性核
				2. 惩罚参数：1
				3. 不敏感参数：输出变量的四分位数范围对标准差的十分之一的估计值
TWP-GPR	充放电压曲线	TWP	F¹, F², F³, F⁴	1. 协方差函数：平方指数核
TWP-GPR	充放电压曲线	TWP	F¹, F², F³, F⁴	2. 噪声标准偏差的初始值：输出变量的标准差除以根号二

准确性指标	拟合模型	稳定性指标
准确性指标	拟合模型	样本标准差SSD	四分位差IQR
RMSE	TWP-SVR	0.0046	0.0061
RMSE	TWP-GPR	0.0051	0.0095
MAE	TWP-SVR	0.0044	0.0054
MAE	TWP-GPR	0.0047	0.0092
MAPE	TWP-SVR	0.6886	0.8410
MAPE	TWP-GPR	0.5625	1.0553

参数	采样周期/ms	误差/%	计算周期/s	更新周期/s	备注
电流	25	±0.2	1	60	1秒钟计算一次平均电流，60秒后上传60个电流数据至服务器
电压	50	±0.3	1	60	1秒钟计算一次平均电压，60秒后上传60个电压数据至服务器
电荷量	25	2	1	60	根据电流采样值，使用安时积分法，1秒钟计算一次电荷量，60秒后计上传60个电荷量数据至服务器
SOC	25	2	1	60	根据电荷量计算荷电状态(SOC)，每次放电初始值为100%
SOH	25	2	1	60	根据电荷量计算SOH，电池组装机第一次放电时为100%