一种基于集成机器学习的液态金属电池快速分选方法

doi:10.19595/j.cnki.1000-6753.tces.221574

Abstract
Figure/Table
References
Related Citation (8)

Download: PDF (1803 KB) HTML (1 KB)
Export: BibTeX | EndNote (RIS)

Abstract Liquid metal battery (LMB) is a newly emerged battery technology for large scale energy storage, which requires massive batteries in group operation. Battery sorting, which could effectively reduce the inconsistency among batteries, is a useful method to significantly improve the performance of energy storage and reduce potential safety risks. However, because of the characteristics of LMB, including large capacity and small internal resistance, existing sorting methods, which mainly focus on lithium ion batteries, cannot meet the precision requirements of liquid metal battery sorting. In addition, these methods also require additional battery testing to obtain input features, thus there is room for further optimization of test time and cost. In order to address these issues, a rapid sorting method for liquid metal batteries based on ensemble learning is proposed. Combining the advantages of different models by adopting feature selection and ensemble learning methods, it precisely predicts the sorting index and conducts battery sorting with satisfactory accuracy.
Firstly, a dataset is constructed based on LMB’s cycling data during activation, which comprises samples in 1D vector form. Secondly, features of the dataset are selected on the basis of a comprehensive method, which takes various aspects into consideration. Thirdly, three different base models are trained and optimized with the help of three-fold validation and grid search optimization, and then two ensemble model, which adopts voting and stacked ensemble method respectively, are trained based on these base models. Fourthly, all the models, including base models and ensemble ones, and a contrast model (neural network) are evaluated on the test set, which demonstrates the superiority of the stacked ensemble model. Finally, battery sorting is conducted on the basis of capacity predictions made by the stacked ensemble model. In this sorting method, the feature selection method reduces the dimension of data and selects the most effective features and ensemble methods integrate advantages of different base models, both of which contribute to the improvement of capacity predictions’ precision and reliability.
Evaluation results on the test set show that the stacked ensemble model has the best performance, with a root mean square error (RMSE) of 0.227 8 A·h, a root mean square percentage error (RMSPE) of 0.455 6% and reliability of 97.42%. In contrast with the mainstream model neural network, its prediction error is reduced by 52.16% and the reliability is improved by 9.10%. Furthermore, compared with models without feature selection, all the models have better overall performance after feature selection, with better precision and higher reliability. Sorting results demonstrate that when the best model, the stacking ensemble model, is applied to LMB sorting, the proposed sorting method achieves both high accuracy and recall rate in battery sorting, which are 96.62% and 93.18% respectively.
The following conclusions can be drawn from the evaluation and sorting results: (1) Compared with existing sorting methods, the proposed method not only saves a lot of time by utilizing data during the activation period instead of data from extra test but also achieves high accuracy and recall rate in battery sorting on the basis of a precise and reliable ensemble model. (2) The comprehensive feature selection method adopted by the proposed sorting method is effective. In a comparison of models without feature selection, models trained after feature selection, including base models and ensemble models, have achieved overall performance improvement with smaller errors and higher reliability. (3) The performance of the proposed ensemble model is not only better than the base model used for ensemble learning, but also superior to the mainstream neural network model on the basis of essential indicators such as model validation score, test set score, and reliability. Therefore, it is a promising model for sorting index prediction in LMB sorting applications.

Key words： Battery sorting liquid metal battery machine learning ensemble learning feature selection

Received: 15 August 2022

PACS:

TM911

	Service

	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	Xia Junyi
	Shi Qionglin
	Jiang Kai
	He Yaling
	Wang Kangli

Cite this article:

Xia Junyi,Shi Qionglin,Jiang Kai等. A Rapid Liquid Metal Battery Sorting Method Based on Ensemble Learning[J]. Transactions of China Electrotechnical Society, 2023, 38(21): 5900-5912.

URL:

https://dgjsxb.ces-transaction.com/EN/10.19595/j.cnki.1000-6753.tces.221574 OR https://dgjsxb.ces-transaction.com/EN/Y2023/V38/I21/5900

[1] 任丽彬, 许寒, 宗军, 等. 大规模储能技术及应用的研究进展[J]. 电源技术, 2018, 42(1): 139-142.
Ren Libin, Xu Han, Zong Jun, et al.Research progress of large-scale energy storage technologies and applications[J]. Chinese Journal of Power Sources, 2018, 42(1): 139-142.
[2] 李培平, 姚伟, 高东学, 等. 基于电化学储能的多馈入直流系统暂态控制及影响因素分析[J]. 电工技术学报, 2021, 36(增刊1): 154-167.
Li Peiping, Yao Wei, Gao Dongxue, et al.Transient control and influencing factors analysis of multi-infeed HVDC system based on electrochemical energy storage[J]. Transactions of China Electrotechnical Society, 2021, 36(S1): 154-167.
[3] 张超, 冯忠楠, 邓少平, 等. 考虑电热混合储能的多能互补协同削峰填谷策略[J]. 电工技术学报, 2021, 36(增刊1): 191-199.
Zhang Chao, Feng Zhongnan, Deng Shaoping, et al.Multi-energy complementary collaborative peak-load shifting strategy based on electro-thermal hybrid energy storage system[J]. Transactions of China Electrotechnical Society, 2021, 36(S1): 191-199.
[4] Wang Kangli, Jiang Kai, Chung B, et al.Lithium-antimony-lead liquid metal battery for grid-level energy storage[J]. Nature, 2014, 514(7522): 348-350.
[5] Kim H, Boysen D A, Newhouse J M, et al.Liquid metal batteries: past, present, and future[J]. Chemical Reviews, 2013, 113(3): 2075-2099.
[6] Zhang Shilin, Liu Ye, Fan Qining, et al.Liquid metal batteries for future energy storage[J]. Energy & Environmental Science, 2021, 14(8): 4177-4202.
[7] Li Haomiao, Yin Huayi, Wang Kangli, et al.Liquid metal electrodes for energy storage batteries[J]. Advanced Energy Materials, 2016, 6(14): 1600483.
[8] Li Haomiao, Wang Kangli, Cheng Shijie, et al.High performance liquid metal battery with environmentally friendly antimony-tin positive electrode[J]. ACS Applied Materials & Interfaces, 2016, 8(20): 12830-12835.
[9] Zhou Hao, Li Haomiao, Gong Qing, et al.A sodium liquid metal battery based on the multi-cationic electrolyte for grid energy storage[J]. Energy Storage Materials, 2022, 50: 572-579.
[10] 李建林, 武亦文, 王楠, 等. 吉瓦级电化学储能电站研究综述及展望[J]. 电力系统自动化, 2021, 45(19): 2-14.
Li Jianlin, Wu Yiwen, Wang Nan, et al.Review and prospect of gigawatt-level electrochemical energy storage power station[J]. Automation of Electric Power Systems, 2021, 45(19): 2-14.
[11] 李建林, 武亦文, 王楠, 等. 吉瓦级电化学储能电站信息架构与安防体系综述[J]. 电力系统自动化, 2021, 45(23): 179-191.
Li Jianlin, Wu Yiwen, Wang Nan, et al.Review of information architecture and security system of gigawatt electrochemical energy storage power station[J]. Automation of Electric Power Systems, 2021, 45(23): 179-191.
[12] 郑岳久, 李家琦, 朱志伟, 等. 基于快速充电曲线的退役锂电池模块快速分选技术[J]. 电网技术, 2020, 44(5): 1664-1673.
Zheng Yuejiu, Li Jiaqi, Zhu Zhiwei, et al.Rapid classification based on fast charging curves for reuse of retired lithium-ion battery modules[J]. Power System Technology, 2020, 44(5): 1664-1673.
[13] Lai Xin, Qiao Dongdong, Zheng Yuejiu, et al.A rapid screening and regrouping approach based on neural networks for large-scale retired lithium-ion cells in second-use applications[J]. Journal of Cleaner Production, 2019, 213: 776-791.
[14] 吕治强, 高仁璟, 黄现国. 基于多核相关向量机优化模型的锂离子电池容量在线估算[J]. 电工技术学报: 2023, 38(7): 1713-1722.
Lü Zhiqiang, Gao Renjing, Huang Xianguo.A Li-Ion Battery Capacity Estimation Method Based on Multi-Kernel Relevance Vector Machine Optimized Model[J]. Transactions of China Electrotechnical Society, 2023, 38(7): 1713-1722.
[15] 杨胜杰, 罗冰洋, 王菁, 等. 基于容量增量曲线峰值区间特征参数的锂离子电池健康状态估算[J]. 电工技术学报, 2021, 36(11): 2277-2287.
Yang Shengjie, Luo Bingyang, Wang Jing, et al.State of health estimation for lithium-ion batteries based on peak region feature parameters of incremental capacity curve[J]. Transactions of China Electrotechnical Society, 2021, 36(11): 2277-2287.
[16] Rastegarpanah A, Hathaway J, Ahmeid M, et al.A rapid neural network-based state of health estimation scheme for screening of end of life electric vehicle batteries[J]. Proceedings of the Institution of Mechanical Engineers, Part I: Journal of Systems and Control Engineering, 2021, 235(3): 330-346.
[17] Zhou Ping, He Zhonglin, Han Tingting, et al.A rapid classification method of the retired LiCo_xNi_yMn₁-_x-_yO₂ batteries for electric vehicles[J]. Energy Reports, 2020, 6: 672-683.
[18] Fang Kaizheng, Chen Shi, Mu Daobin, et al.Investigation of nickel-metal hydride battery sorting based on charging thermal behavior[J]. Journal of Power Sources, 2013, 224: 120-124.
[19] Garg A, Yun Liu, Gao Liang, et al.Development of recycling strategy for large stacked systems: experimental and machine learning approach to form reuse battery packs for secondary applications[J]. Journal of Cleaner Production, 2020, 275: 124152.
[20] He Fengxian, Shen W X, Song Qiang, et al.Clustering LiFePO₄ cells for battery pack based on neural network in EVs[C]//2014 IEEE Conference and Expo Transportation Electrification Asia-Pacific (ITEC Asia-Pacific), Beijing, China, 2014: 1-5.
[21] Lü Chao, Song Yankong, Wang Lixin, et al.A new method for lithium-ion battery uniformity sorting based on internal criteria[J]. Journal of Energy Storage, 2019, 25: 100885.
[22] Ran Aihua, Zhou Zihao, Chen Shuxiao, et al.Data-driven fast clustering of second-life lithium-ion battery: mechanism and algorithm[J].Advanced Theory and Simulations, 2020. 3(8): 2000109.
[23] 来鑫, 陈权威, 邓聪, 等. 一种基于电化学阻抗谱的大规模退役锂离子电池的软聚类方法[J]. 电工技术学报, 2022, 37(23): 6054-6064.
Lai Xin, Chen Quanwei, Deng Cong, et al.A soft clustering method for the large-scale retired lithium-ion batteries based on electrochemical impedance spectroscopy[J]. Transactions of China Electrotechnical Society, 2022, 37(23): 6054-6064.
[24] 张娥, 徐成, 王晟, 等. 基于模糊逻辑控制器的液态金属电池组两级均衡系统[J]. 中国电机工程学报, 2020, 40(12): 4024-4033.
Zhang E, Xu Cheng, Wang Sheng, et al.Two-stage equalizing system of liquid metal batteries based on fuzzy logic controller[J]. Proceedings of the CSEE, 2020, 40(12): 4024-4033.
[25] 李浩秒, 周浩, 王康丽, 等. 液态金属电极的电化学储能应用[J]. 电化学, 2020, 26(5): 663-682.
Li Haomiao, Zhou Hao, Wang Kangli, et al.Liquid metal electrodes for electrochemical energy storage technologies[J]. Journal of Electrochemistry, 2020, 26(5): 663-682.
[26] 蒋凯, 李浩秒, 李威, 等. 几类面向电网的储能电池介绍[J]. 电力系统自动化, 2013, 37(1): 47-53.
Jiang Kai, Li Haomiao, Li Wei, et al.On several battery technologies for power grids[J]. Automation of Electric Power Systems, 2013, 37(1): 47-53.
[27] Yan Shuai, Zhou Xianbo, Li Haomiao, et al.Utilizing in situ alloying reaction to achieve the self-healing, high energy density and cost-effective Li||Sb liquid metal battery[J]. Journal of Power Sources, 2021, 514: 230578.
[28] Zhou Xianbo, Zhou Hao, Yan Shuai, et al.Increasing the actual energy density of Sb-based liquid metal battery[J]. Journal of Power Sources, 2022, 534: 231428.
[29] 王大磊, 王康丽, 程时杰, 等. 液态金属电池储能特性建模及荷电状态估计[J]. 中国电机工程学报, 2017, 37(8): 2253-2261.
Wang Dalei, Wang Kangli, Cheng Shijie, et al.Modeling of energy storage properties and SOC estimation for liquid metal batteries[J]. Proceedings of the CSEE, 2017, 37(8): 2253-2261.
[30] Liu Guoan, Xu Cheng, Li Haomiao, et al.State of charge and online model parameters co-estimation for liquid metal batteries[J]. Applied Energy, 2019, 250: 677-684.
[31] Smola A J, Schölkopf B.A tutorial on support vector regression[J]. Statistics and Computing, 2004, 14(3): 199-222.
[32] Vapnik V N.The Nature of Statistical Learning Theory[M]. New York: Springer, 1995.
[33] Chen Tianqi, Guestrin C.XGBoost: a scalable tree boosting system[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, USA, 2016: 785-794.
[34] Breiman L.Random forests[J]. Machine Learning, 2001, 45(1): 5-32.
[35] Wolpert D H.Stacked generalization[J]. Neural Networks, 1992, 5(2): 241-259.
[36] Pedregosa F, Varoquaux G, Gramfort A, et al.Scikit-learn: machine learning in python[J]. Journal of Machine Learning Research, 2012, 12: 2825-2830.
[37] Cover T M, Thomas J A.Elements of Information Theory[M]. 2nd ed. New York: Wiley-Interscience, 2005.
[38] Ferri F J, Pudil P, Hatef M, et al.Comparative study of techniques for large-scale feature selection[J]. Machine Intelligence and Pattern Recognition, 1994, 16: 403-413.
[39] Geurts P, Ernst D, Wehenkel L.Extremely randomized trees[J]. Machine Learning, 2006, 63(1): 3-42.