基于自适应灾变遗传-循环神经网络的锂离子电池SOC估计*

doi:10.11985/2022.01.012

摘要/Abstract

摘要：

锂离子荷电状态(State of charge,SOC)的精准估计是锂离子电池安全稳定运行的基础。传统的误差反向传播(Back propagation,BP)神经网络估计SOC的精度不高,而循环神经网络(Recurrent neural network,RNN)也容易陷入局部最优。针对这些问题,提出了自适应灾变遗传-循环神经网络(ACGA-RNN)联合算法,将自适应灾变遗传算法(Adaptive cataclysm genetic algorithm,ACGA)用于优化RNN的初始权值和阈值,提高了最优权值和阈值的全局搜索能力,从而有效提升锂离子电池SOC的估计精度。基于锂离子电池充放电的试验数据,将所提ACGA-RNN联合算法与RNN、GA-RNN算法分别用于锂离子电池的SOC估计。测试结果显示,相较于传统的RNN算法与GA-RNN算法,提出的ACGA-RNN联合算法获得了最佳的SOC估计精度,在DST工况下的估计平均绝对误差为1.74%,低于传统RNN和GA-RNN的估计精度3.68%和2.49%;另外,在45 ℃和0 ℃条件下,ACGA-RNN联合算法估计的平均绝对值误差分别为1.75%和2.05%,符合国家标准要求。因此,提出的ACGA-RNN联合算法在锂电池的SOC估计方面具有良好的应用价值。

关键词: 锂离子电池, 荷电状态, 循环神经网络, 自适应灾变遗传算法

Abstract:

Accurate estimation of the state of charge (SOC) of lithium batteries is the basis for safe and stable operation of lithium batteries. The traditional back propagation (BP) neural network has low accuracy in estimating SOC, and the recurrent neural network (RNN) is easy to fall into local optimization. To work out these problems, a recurrent neural network (ACGA-RNN) combined algorithm based on an adaptive catastrophe genetic algorithm is proposed. The adaptive catastrophe genetic algorithm (ACGA) is utilized to optimize the initial weight and threshold of RNN and improve the global search ability of the optimal value, which effectively improves the estimation preciseness of the state of charge of lithium battery. Based on the experimental data from the charge and discharge test, the proposed ACGA-RNN combined algorithm, RNN algorithm, and GA-RNN algorithm are applied for SOC estimation of the lithium-ion battery respectively. The experimental results show that the estimation mean absolute error of ACGA-RNN combined algorithm obtains the best accuracy of SOC estimation compared with the traditional RNN and GA-RNN algorithms. The mean absolute error under the DST condition is 1.74%, which is lower than the estimation accuracy of traditional RNN (3.68%) and GA-RNN (2.49%). In addition, at 45 ℃ and 0 ℃, the mean absolute error of the ACGA-RNN combined algorithm is 1.75% and 2.05%, respectively, which meets the requirements of national standards. Therefore, the proposed ACGA-RNN combined algorithm has a good application value in SOC estimation of lithium batteries.

Key words: Lithium battery, state of charge, recurrent neural network, adaptive cataclysm genetic algorithm

中图分类号:

TM912

陈诚, 皮志勇, 赵英龙, 廖玄, 张明敏, 李勇. 基于自适应灾变遗传-循环神经网络的锂离子电池SOC估计^*[J]. 电气工程学报, 2022, 17(1): 86-94.

CHEN Cheng, PI Zhiyong, ZHAO Yinglong, LIAO Xuan, ZHANG Mingmin, LI Yong. State of Charge Estimation with Adaptive Cataclysm Genetic Algorithm-recurrent Neural Network for Li-ion Batteries[J]. Journal of Electrical Engineering, 2022, 17(1): 86-94.

图/表 15

图1

图2

表1

图3

表2

表3

图4

图5

表4

图6

图7

表5

图8

图9

表6

参考文献 29

[1]	李晖, 刘栋, 姚丹阳. 面向碳达峰碳中和目标的我国电力系统发展研判[J]. 中国电机工程学报, 2021, 29(7):1-13.
	LI Hui, LIU Dong, YAO Danyang. Analysis and reflection on the development of power system towards the goal of carbon emission peak and carbon neutrality[J]. Proceedings of the Chinese Society of Electrical Engineering, 2021, 29(7):1-13.
[2]	李建林, 武亦文, 吕超, 等. 碳达峰背景下退役锂电池梯次利用电热管理技术[J]. 高电压技术, 2021, 29(7):1-16.
	LI Jianlin, WU Yiwen, LÜ Chao, et al. Electro&thermal management technology for cascade utilization of retired lithium batteries under the background of carbon peak[J]. High Voltage Engineering, 2021, 29(7):1-16.
[3]	姜媛媛, 曾文文, 沈静静, 等. 基于凸优化-寿命参数退化机理模型的锂离子电池剩余使用寿命预测[J]. 电力系统及其自动化学报, 2019, 31(3):23-28.
	JIANG Yuanyuan, ZENG Wenwen, SHEN Jingjing, et al. Prediction of remaining useful life of lithium-ion battery based on convex optimization life parameter degradation mechanism model[J]. Proceedings of the CSU-EPSA, 2019, 31(3):23-28.
[4]	赖春艳, 陈宏, 倪嘉茜, 等. 锂离子电池储能技术在电力能源中的应用模式与发展趋势[J]. 上海电力大学学报, 2021, 37(4):380-384.
	LAI Chunyan, CHEN Hong, NI Jiaqian, et al. Application mode and development trend of lithium ion battery energy storage technology in power energy[J]. Journal of Shanghai Electric Power University, 2021, 37(4):380-384.
[5]	杨潇, 王顺利, 徐文华, 等. 基于LM-RLS和渐消因子EKF算法的锂电池SOC估计[J]. 控制工程, 2021, 29(7):1-7.
	YANG Xiao, WANG Shunli, XU Wenhua, et al. SOC estimation of lithium battery based on LM-RLS and fading factor EKF algorithm[J]. Control Engineering, 2021, 29(7):1-7.
[6]	吕甜, 张雪霞. 基于在线参数辨识和ASR-UKF的锂离子电容SOC估计[J]. 电源技术, 2021, 45(1):27-30,55.
	LÜ Tian, ZHANG Xuexia. SOC estimation of lithium-ion capacitor based on online parameter identification and ASR-UKF[J]. Chinese Journal of Power Sources, 2021, 45(1):27-30,55.
[7]	张武, 孙士山, 张家福. 基于自适应无迹卡尔曼滤波的动力电池SOC估计[J]. 电源技术, 2021, 45(1):14-17.
	ZHANG Wu, SUN Shishan, ZHANG Jiafu. SOC estimation of power battery based on adaptive unscented Kalman filter[J]. Chinese Journal of Power Sources, 2021, 45(1):14-17.
[8]	李超然, 肖飞, 樊亚翔, 等. 基于高斯过程回归的锂电池SOC估算方法[J]. 海军工程大学学报, 2021, 33(1):55-59.
	LI Chaoran, XIAO Fei, FAN Yaxiang, et al. Electro&thermal SOC estimation method of lithium battery based on Gaussian process regression[J]. Journal of Navy Engineering University, 2021, 33(1):55-59.
[9]	陈霖华, 陈剑, 徐志强, 等. 基于实时电路模型的储能系统锂离子电池状态估算[J]. 中南大学学报, 2021, 52(2):458-464.
	CHEN Linhua, CHEN Jian, XU Zhiqiang, et al. State estimation of lithium ion battery in energy storage system based on real-time circuit model[J]. Journal of Central South University, 2021, 52(2):458-464.
[10]	孙鹏宇, 李建良, 陶知非, 等. 动态工况电池在线参数辨识及SOC估计研究[J]. 电子测量与仪器学报, 2021, 35(1):10-17.
	SUN Pengyu, LI Jianliang, TAO Zhifei, et al. On-line parameter identification and SOC estimation of battery under dynamic conditions[J]. Journal of Electronic Measurement and Instrument, 2021, 35(1):10-17.
[11]	谈发明, 赵俊杰, 李秋烨. 基于简化滞回OCV模型的锂电池SOC自适应估计策略[J]. 中国电机工程学报, 2021, 41(2):703-715.
	TAN Faming, ZHAO Junjie, LI Qiuye. SOC adaptive estimation strategy of lithium battery based on simplified hysteresis OCV model[J]. Proceedings of the Chinese Society of Electrical Engineering, 2021, 41(2):703-715.
[12]	周娟, 孙啸, 刘凯, 等. 联合扩展卡尔曼滤波的滑模观测器SOC估算算法研究[J]. 中国电机工程学报, 2021, 41(2):692-703.
	ZHOU Juan, SUN Xiao, LIU Kai, et al. Research on SOC estimation algorithm of sliding mode observer combined with extended Kalman filter[J]. Proceedings of the Chinese Society of Electrical Engineering, 2021, 41(2):692-703.
[13]	王晓兰, 靳皓晴, 刘祥远. 基于融合模型的锂离子电池荷电状态在线估计[J]. 工程科学学报, 2020, 42(9):1200-1208.
	WANG Xiaolan, JIN Haoqing, LIU Xiangyuan. Online estimation of charge state of lithium ion battery based on fusion model[J]. Journal of Engineering Science, 2020, 42(9):1200-1208.
[14]	李静, 杨艳, 秦建鑫, 等. 锂电池分数阶建模与荷电状态估计研究[J]. 电源技术, 2020, 44(7):983-985,994.
	LI Jing, YANG Yan, QIN Jianxin, et al. Fractional modeling and state of charge estimation of lithium battery[J]. Chinese Journal of Power Sources, 2020, 44(7):983-985,994.
[15]	秦鹏, 王振新, 康健强, 等. 实时辨识锂离子电池参数并基于改进AEKF估算SOC[J]. 电子测量技术, 2020, 43(10):30-35.
	QIN Peng, WANG Zhenxin, KANG Jianqiang, et al. Real-time identification of lithium ion battery parameters and SOC estimation based on improved AEKF[J]. Electronic Measurement Techniques, 2020, 43(10):30-35.
[16]	孙硕, 孙俊忠, 周智勇, 等. 基于MIV-GA-BP神经网络的铅酸蓄电池SOC预测[J]. 电源技术, 2021, 45(2):228-231.
	SUN Shuo, SUN Junzhong, ZHOU Zhiyong, et al. SOC prediction of lead-acid battery based on MIV-GA-BP neural network[J]. Chinese Journal of Power Sources, 2021, 45(2):228-231.
[17]	姚芳, 张楠, 黄凯. 估算锂电池SOC的基于LM的BP神经网络算法[J]. 电源技术, 2019, 43(9):1453-1457.
	YAO Fang, ZHANG Nan, HUANG Kai. LM-based BP neural network algorithm for estimating SOC of lithium battery[J]. Chinese Journal of Power Sources, 2019, 43(9):1453-1457.
[18]	赵钢, 朱芳欣, 窦汝振. 基于PSO-BP的电动汽车锂离子电池SOC估算[J]. 电源技术, 2018, 42(9):1318-1320.
	ZHAO Gang, ZHU Fangxin, DOU Ruzhen. SOC estimation of electric vehicle lithium-ion battery based on PSO-BP[J]. Chinese Journal of Power Sources, 2018, 42(9):1318-1320.
[19]	BOCKRATH S, ROSSKOPF A, KOFFEL S, et al. State of charge estimation using recurrent neural networks with long short-term memory for lithium-ion batteries[C]// IECON 2019-45th Annual Conference of the IEEE Industrial Electronics Society, 2019, 9(10):2507-2511.
[20]	YANG Fangfang, LI Weihua, LI Chuan, et al. State-of-charge estimation of lithium-ion batteries based on gated recurrent neural network[J]. Energy, 2019, 15(5):66-75.
[21]	林海飞, 高帆, 严敏, 等. 煤层瓦斯含量PSO-BP神经网络预测模型及其应用[J]. 中国安全科学学报, 2020, 30(9):80-87.
	LIN Haifei, GAO Fan, YAN Min, et al. PSO-BP neural network prediction model of coal seam gas content and its application[J]. China Safety Science, 2020, 30(9):80-87.
[22]	陈颖, 黄凯, 丁恒, 等. 基于子种群自适应思维进化-BP神经网络的锂离子电池SOC估计[J]. 电源学报, 2021, 29(7):1-15.
	CHEN Ying, HUANG Kai, DING Heng, et al. SOC estimation of lithium ion battery based on subpopulation adaptive thinking evolution-BP neural network[J]. Journal of Power Supply, 2021, 29(7):1-15.
[23]	颜建国, 郑书闽, 郭鹏程, 等. 基于GA-BP神经网络的超临界CO_(2) 传热特性预测研究[J]. 化工学报, 2021, 29(7):1-16.
	YAN Jianguo, ZHENG Shumin, GUO Pengcheng, et al. Prediction of supercritical CO_(2) heat transfer characteristics based on GA-BP neural network[J]. Journal of Chemical Industry and Engineering China, 2021, 29(7):1-16.
[24]	幸火林, 徐安军, 吴双平, 等. 基于改进遗传算法的热轧批量计划优化模型[J]. 中国冶金, 2021, 31(5):47-53.
	XING Huolin, XU Anjun, WU Shuangping, et al. Hot rolling batch planning optimization model based on improved genetic algorithm[J]. China Metallurgy, 2021, 31(5):47-53.
[25]	LIU Yaoyao, ZHANG Yixin, LIU Xiaoya. Application of embedded computer and improved genetic algorithm in the strategy of community of human destiny:The development of artificial intelligence in the context of Covid-19[J]. Journal of Ambient Intelligence and Humanized Computing, 2021, 9(4):1-13. doi: 10.1007/s12652-017-0523-3
[26]	HAMED K, MOHAMMAD M M, MOHAMMAD R, et al. The integrated production-distribution scheduling in parallel machine environment by using improved genetic algorithms[J]. Journal of Industrial and Production Engineering, 2021, 38(3):157-170. doi: 10.1080/21681015.2020.1848930
[27]	王策, 董兆伟, 孙立辉, 等. 新型灾变自适应遗传算法及其应用[J]. 计算机系统应用, 2019, 28(9):278-283.
	WANG Ce, DONG Zhaowei, SUN Lihui, et al. New catastrophe adaptive genetic algorithm and application[J]. Computer Systems & Applications, 2019, 28(9):278-283.
[28]	University of Maryland. Battery research data[EB/OL]. [2021-07-08]. https://calce.umd.edu/data .
[29]	中华人民共和国国家市场监督管理总局, 中国国家标准化管理委员会. GB/T 38611-2020 电动汽车用电池管理系统技术条件[S]. 北京: 中国标准出版社, 2020.
	State Administration for Market Regulation, Inspection and Quarantine of the People’s Republic of China,Standardization Administration of the People’s Republic of China. GB/T 38611-2020 Technical specifications of battery management system for electric vehicles[S]. Beijing: Standards Press of China, 2020.

RNN参数名称	参数值
状态-状态权值	[ 0.200 2 1.829 1.466, -0.270 1 0.324 2, -0.172 8 -0.888 4 1.295] [0.522 4 -1.576 0.197 5 0.269 1, 0.647 7 -1.912 0.814 8 -1.548] [ 1.937 0.318 9 0.071 07 -1.05 0.712 9 1.26 -0.080 06 -0.021 78] [0.610 6 0.826 7 -0.340 6 -1.0 1.646, -1.293 0.826 3 -1.805] [ 0.401 8 0.327 8 1.096 -1.124 1.637 -1.338 -1.78 1.724] [ 1.68 -1.141 0.616 1 1.515 -1.039, -1.32 -0.253 1 1.066] [0.284 9 0.518 3 0.526 4 -0.519 7, 1.922 0.770 5 0.599 9 0.803] [-0.663 8 -1.078 -1.376 0.288 3, -1.273 -1.171 0.204 7 -0.824 5]
状态-输出权值	[1.724, 0.565 1] [-0.702 1, -1.688] [1.837, -1.119] [-1.723, -1.774] [1.467, -1.367] [0.778 6, 0.771 5] [-1.424, 0.975] [0.819 9, 0.227 3]
状态-输入权值	[-0.602 7, -1.33, -0.019 56, 1.213, 0.595 4, 0.754 2, 0.652 2, 0.931 7]
状态阈值	[ 1.378, 0.500 7, -1.28, 1.109, -1.345, 1.64, -1.695, -0.635 9]^T
输出阈值	0.322 5

超参数	数值
训练次数	3 000
隐藏层个数	8
序列长度	3
初始学习率	0.001
优化器	Adam
激活函数	Leaky ReLU

参数	数值
品牌	三星
名称	INR 18650-20R
寿命	500~800循环
额定容量/(mA·h)	2 000
额定电压/V	3.6
充电截止电压/V	4.2
放电截止电压/V	2.5

估计方法	RMSE(%)	MAE(%)
RNN	4.76	4.21
GA-RNN	3.32	2.75
ACGA-RNN	1.97	1.72

温度/℃	RMSE(%)	MAE(%)
0	2.23	2.05
25	1.98	1.48
45	2.12	1.72