基于Bi-LSTM/Bi-GRU循环神经网络的锂电池SOC估计

doi:10.19799/j.cnki.2095-4239.2021.0051

摘要/Abstract

摘要：

锂电池的荷电状态(state-of-charge， SOC)涉及的物理特性或电化学特性高度复杂，其值一般难以直接测量，基于深度神经网络等新方法的SOC估计近期为相关研究者所关注。为进一步提升SOC估计性能，有效捕获锂电池SOC的动态物理特性，缓解深度神经网络模型容易发生的梯度消失与梯度爆炸等问题，本文引入双向学习策略，基于双向长短期记忆循环神经网络(bidirectional long short-term memory， Bi-LSTM)以及双向门控循环单元网络(bidirectional gated recurrent unit， Bi-GRU)估计锂电池的SOC取值。双向循环神经网络SOC估计模型由输入层、隐藏层和输出层组成。输入层输入电池电压、电流与温度序列；隐藏层在正向LSTM/GRU 层的基础上增加反向LSTM/GRU 层，引入逆序信息，基于输入序列上下文所含信息整体上学习、表征电池特性序列与SOC序列之间的内在关联；输出层输出模型的估计值。所拟模型使用Python 语言结合TensorFlow 后端在Keras 框架中实现，并基于基准数据集在3种温度条件下结合多种工况进行性能分析。结果表明，双向学习策略能有效提升锂电池SOC的估计性能，较之单向学习模型具有更高的估计精度与鲁棒性。与构造电池等效模型等方法的思路不同，所拟方法基于数据驱动学习锂电池SOC的非线性特性，将易于观测的锂电池特性序列数据映射为待估计的SOC取值，为锂电池SOC估计提供了可能的新思路。

关键词: 锂电池, SOC估计, Bi-LSTM网络, Bi-GRU网络

Abstract:

The direct measurement of the state of charge (SOC) for lithium-ion batteries, whose physical characteristics or the electrochemical characteristics are highly complex, presents difficulty. Methods, such as deep neural networks, have recently received wide attention and have been exploited to estimate battery SOC. To improve the estimation performance of SOC, capture accurately the interior dynamic characteristics of the batteries, and alleviate the vanishing and exploding gradient problem in neural networks, in this paper, we introduce a bidirectional learning strategy and develop two specific methods, i.e., the estimation of the SOC with the bidirectional long short-term memory (Bi-LSTM) networks and the bidirectional gated recurrent unit (Bi-GRU) networks. Both the proposed bidirectional networks consist of three parts: the input, hidden, and output layers. The input layer accepts the sequence of measurements, such as the voltages, currents, and temperatures. The hidden layer consists of two LSTM/GRU sub-layers; one processes the forward information, and the other processes the backward information to learn the input-output sequence mapping form the context of the sequence. The output layer outputs the learned SOC. Python, TensorFlow, and Keras are used to develop the concrete models, and the resultant ones are tested and analyzed on the benchmark data sets, where three temperature and nine working conditions are considered. The results show that the proposed bidirectional learning methods can outperform their competitors, with higher accuracy and better robustness. Being different from the idea of constructing battery equivalent models, the proposed method is data-driven and intends to estimate SOC with easily obtainable measurements, such as the voltages, currents, and temperatures, and provides a potential method for SOC estimation.

Key words: lithium-ion battery, SOC estimation, Bi-LSTM network, Bi-GRU network

中图分类号:

TM 912

朱元富, 贺文武, 李建兴, 李有财, 李培强. 基于Bi-LSTM/Bi-GRU循环神经网络的锂电池SOC估计[J]. 储能科学与技术, 2021, 10(3): 1163-1176.

Yuanfu ZHU, Wenwu HE, Jianxing LI, Youcai LI, Peiqiang LI. SOC estimation for Li-ion batteries based on Bi-LSTM and Bi-GRU[J]. Energy Storage Science and Technology, 2021, 10(3): 1163-1176.

图/表 18

图1

图2

图3

图4

图5

图6

表1

图7

图8

表2

图9

表3

图10

图11

图12

图13

图14

图15

参考文献 24

1	GAO W, ZHANG X, ZHENG X, et al. Lithium carbonate recovery from cathode scrap of spent lithium-ion battery: A closed-loop process[J]. Environmental Science & Technology, 2017, 51(3): 1662-1669.
2	SHRIVASTAVA P, SOON T K, IDRIS M Y I B, et al. Overview of model-based online state-of-charge estimation using Kalman filter family for lithium-ion batteries[J]. Renewable and Sustainable Energy Reviews, 2019, 113: doi: 10.1016/j.rser.2019.06.040.
3	CHEMALI E, PREINDL M, MALYSZ P, et al. Electrochemical and electrostatic energy storage and management systems for electric drive vehicles: State-of-the-art review and future trends[J]. IEEE Journal of Emerging & Selected Topics in Power Electronics, 2016, 4(3): 1117-1134.
4	李超然, 肖飞, 樊亚翔, 等. 基于门控循环单元神经网络和Huber-M估计鲁棒卡尔曼滤波融合方法的锂离子电池荷电状态估算方法[J]. 电工技术学报, 2020, 35(9): 2051-2062.
	LI C R, XIAO F, FAN Y X, et al. A hybrid approach to lithium-ion battery SOC estimation based on recurrent neural network with gated recurrent unit and Huber-M robust Kalman filter[J]. Transactions of China Electrotechnical Society, 2020, 35(9): 2051-2062.
5	丁镇涛, 邓涛, 李志飞, 等. 基于安时积分和无迹卡尔曼滤波的锂电池SOC估算方法研究[J]. 中国机械工程, 2020, 31(15): 1823-1830.
	DING Z T, DENG T, LI Z F, et al. SOC estimation of lithium-ion battery based on ampere hour integral and unscented Kalman filter[J]. China Mechanical Engineering, 2020, 31(15): 1823-1830.
6	LIN C, YU Q Q, XIONG R, et al. A study on the impact of open circuit voltage tests on state of charge estimation for lithium-ion batteries[J]. Applied Energy, 2017, 205: 892-902.
7	SKOOG S, DAVID S. Parameterization of linear equivalent circuit models over wide temperature and SOC spans for automotive lithium-ion cells using electrochemical impedance spectroscopy[J]. Journal of Energy Storage, 2017, 14: 39-48.
8	RAMADAN H S, BECHERIF M, CLAUDE F. Extended Kalman filter for accurate state of charge estimation of lithium-based batteries: A comparative analysis[J]. International Journal of Hydrogen Energy, 2017, 42(48): 29033-29046.
9	张远进, 吴华伟, 叶从进. 基于AUKF-BP神经网络的锂电池SOC估算[J]. 储能科学与技术, 2021, 10(1): 237-241.
	ZHANG Y J, WU H W, YE C J. Estimation of the SOC of a battery based on the AUKF-BP algorithm[J]. Energy Storage Science and Technology, 2021, 10(1): 237-241.
10	BACCOUCHE I, MANAI B, AMARA N E B. SOC estimation of LFP battery based on EKF observer and a full polynomial parameters-model[C]//2020 IEEE 91st Vehicular Technology Conference, 2020.
11	YE M, GUO H, XIONG R, et al. A double-scale and adaptive particle filter-based online parameter and state of charge estimation method for lithium-ion batteries[J]. Energy, 2018, 144: 789-799.
12	NING B, XU J, CAO B G, et al. A sliding mode observer SOC estimation method based on parameter adaptive battery model[J]. Energy Procedia, 2016, 88: 619-626.
13	成文晶, 潘庭龙. 基于分布估计算法LSSVM的锂电池SOC预测[J]. 储能科学与技术, 2020, 9(6): 1948-1953.
	CHENG W J, PAN T L. Prediction for SOC of lithium-ion batteries by estimating the distribution algorithm with LSSVM[J]. Energy Storage Science and Technology, 2020, 9(6): 1948-1953.
14	GUO Y F, ZHAO Z S, HUANG L M. SOC estimation of lithium battery based on improved BP neural network[J]. Energy Procedia, 2017, 105: 4153-4158.
15	CHEMALI E, KOLLMEYER P J, PREINDL M, et al. Long short-term memory-networks for accurate state of charge estimation of Li-ion batteries[J]. IEEE Transactions on Industrial Electronics, 2017, 65(8): 6730-6739.
16	CHEN C, XIONG R, YANG R X, et al. State-of-charge estimation of lithium-ion battery using an improved neural network model and extended Kalman filter[J]. Journal of Cleaner Production, 2019, 234: 1153-1164.
17	ELMAN J L. Finding structure in time[J]. Cognitive Science, 1990, 14(2): 179-211.
18	HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735-1780.
19	XU C Y, SHEN J Z, DU X, et al. An intrusion detection system using a deep neural network with gated recurrent units[J]. IEEE Access, 2018, 6: 48697-48707.
20	ARKHIPENKO K, KOZLOV I, TROFIMOVICH J, et al. Comparison of neural network architectures for sentiment analysis of Russian tweets[C]//Computational Linguistics and Intellectual Technologies: Proceedings of the International Conference Dialogue, 2016.
21	GRAVES A, SCHMIDHUBER J. Framewise phoneme classification with bidirectional LSTM and other neural network architectures[J]. Neural Networks, 2005, 18(5/6): 602-610.
22	LIU F G, ZHENG J Z, ZHENG L L, et al. Combining attention-based bidirectional gated recurrent neural network and two-dimensional convolutional neural network for document-level sentiment classification[J]. Neurocomputing, 2020, 371: 39-50.
23	CHEMALI E, KOLLMEYER P J, PREINDL M, et al. State-of-charge estimation of Li-ion batteries using deep neural networks: A machine learning approach[J]. Journal of Power Sources, 2018, 400: 242-255.
24	LI C R, XIAO F, FAN Y X. An approach to state of charge estimation of lithium-ion batteries based on recurrent neural networks with gated recurrent unit[J]. Energies, 2019, 12(9): doi: 10.3390/en12091592.

参数指标	Bi-LSTM/Bi-GRU网络	LSTM/ GRU网络
输入层节点	3	3
神经网络层节点	250	1000
全连接层节点	50	50
输出层节点	1	1
损失函数	MAE	MAE
自适应优化器	Adam	Adam
评价函数	MAE/MAX	MAE/MAX
时间步长	1500	1000
训练迭代次数	100	100

温度/℃	网络结构	HWFET		LA92		US06
温度/℃	网络结构	MAE/%	MAX/%	MAE/%	MAX/%	MAE/%	MAX/%
0	LSTM	3.041	6.711	2.255	9.384	1.496	9.811
	GRU	1.991	3.804	1.086	2.431	0.861	5.542
	Bi-LSTM	1.371	4.024	0.786	4.536	0.694	4.404
	Bi-GRU	0.688	3.197	0.568	4.534	0.596	4.598
10	LSTM	2.176	6.101	0.968	9.069	1.267	8.131
	GRU	1.659	5.624	0.791	7.754	1.169	6.407
	Bi-LSTM	1.282	4.737	0.611	6.742	0.679	5.165
	Bi-GRU	0.952	3.431	0.471	2.558	0.500	3.100
25	LSTM	1.765	4.269	0.791	3.461	1.187	5.621
	GRU	1.149	3.327	0.431	3.519	0.614	3.443
	Bi-LSTM	0.637	2.825	0.375	2.754	0.472	2.731
	Bi-GRU	0.689	2.649	0.309	2.215	0.325	1.696
Total	LSTM	2.327	6.711	1.338	9.384	1.317	9.811
	GRU	1.600	5.624	0.769	7.754	0.881	6.407
	Bi-LSTM	1.097	4.737	0.591	6.742	0.615	5.165
	Bi-GRU	0.776	3.431	0.449	4.534	0.474	4.598

指标	LSTM	GRU	Bi-LSTM	Bi-GRU	Average
MAE/%	1.661	1.083	0.768	0.566	—
Single/%	—	-34.80	-29.09	-26.30	-30.06
Double/%	—	—	-53.76	-47.74	-50.75
MAX/%	9.811	7.754	6.742	4.598	—
Single/%	—	-20.97	-13.05	-31.80	-21.94
Double/%	—	—	-31.28	-40.70	-35.99

[1]	时雨, 张忠, 杨晶莹, 钱薇, 李昊, 赵祥, 杨欣桐. 储能电池系统提供AGC调频的机会成本建模与市场策略[J]. 储能科学与技术, 2022, 11(7): 2366-2373.
[2]	元佳宇, 李昕光, 王文超, 付程阔. 考虑质量流量的电池组蛇形冷却结构仿真[J]. 储能科学与技术, 2022, 11(7): 2274-2281.
[3]	黄鹏, 聂枝根, 陈峥, 舒星, 沈世全, 杨继鹏, 申江卫. 基于优化Elman神经网络的锂电池容量预测[J]. 储能科学与技术, 2022, 11(7): 2282-2294.
[4]	张肖洒, 王宏源, 李振彪, 夏志美. 废旧磷酸铁锂电池电极材料的硫酸化焙烧-水浸新工艺[J]. 储能科学与技术, 2022, 11(7): 2066-2074.
[5]	申晓宇, 岑官骏, 乔荣涵, 朱璟, 季洪祥, 田孟羽, 金周, 闫勇, 武怿达, 詹元杰, 俞海龙, 贲留斌, 刘燕燕, 黄学杰. 锂电池百篇论文点评（2022.4.1—2022.5.31）[J]. 储能科学与技术, 2022, 11(7): 2007-2022.
[6]	周伟东, 黄秋, 谢晓新, 陈科君, 李薇, 邱介山. 固态锂电池聚合物电解质研究进展[J]. 储能科学与技术, 2022, 11(6): 1788-1805.
[7]	乔荣涵, 岑官骏, 申晓宇, 田孟羽, 季洪祥, 田丰, 起文斌, 金周, 武怿达, 詹元杰, 闫勇, 贲留斌, 俞海龙, 刘燕燕, 黄学杰. 锂电池百篇论文点评（2022.2.1—2022.3.31）[J]. 储能科学与技术, 2022, 11(5): 1289-1304.
[8]	王苏杭, 李建林, 李雅欣, 熊俊杰, 曾伟. 锂离子电池系统低温充电策略[J]. 储能科学与技术, 2022, 11(5): 1537-1542.
[9]	汪红辉, 吴泽钦, 储德韧. 轻度过放模式下钛酸锂电池性能及热安全性[J]. 储能科学与技术, 2022, 11(5): 1305-1313.
[10]	郑征, 王肖帅, 李斌, 黄涛, 李佩柯. 基于三绕组变压器的锂电池组自适应交错控制均衡方案[J]. 储能科学与技术, 2022, 11(4): 1131-1140.
[11]	王星星, 宋子钰, 吴浩, 冯文芳, 周志彬, 张恒. 固态聚合物电解质导电锂盐的研究进展[J]. 储能科学与技术, 2022, 11(4): 1226-1235.
[12]	岑官骏, 朱璟, 乔荣涵, 申晓宇, 季洪祥, 田孟羽, 田丰, 金周, 闫勇, 武怿达, 詹元杰, 俞海龙, 贲留斌, 刘燕燕, 黄学杰. 锂电池百篇论文点评（2021.12.1—2022.1.31）[J]. 储能科学与技术, 2022, 11(3): 1077-1092.
[13]	陈博文, 崔瑞广, 沈炎宾, 陈立桅. 杨氏模量微观表征新方法在锂电池中的应用[J]. 储能科学与技术, 2022, 11(3): 991-999.
[14]	翁素婷, 刘泽鹏, 杨高靖, 张思蒙, 张啸, 方遒, 李叶晶, 王兆翔, 王雪锋, 陈立泉. 冷冻电镜表征锂电池中的辐照敏感材料[J]. 储能科学与技术, 2022, 11(3): 760-780.
[15]	余春林, 陈旭东, 宫川敏夫, 孙辉, 张兴旺, 童莉葛. 特殊结构前驱体对锂电池三元正极材料性能的提升[J]. 储能科学与技术, 2022, 11(3): 1000-1007.