作者：李泓沛, 刘桂雄, 邓威

作者单位：华南理工大学机械与汽车工程学院，广东 广州 510640

关键词：动力电池;荷电状态;长短期记忆;无迹卡尔曼滤波;算法融合

摘要：

为提高动力电池荷电状态(state of charge, SOC)估算准确性、稳定性，该文提出一种基于LSTM+UKF(long short term memory+unscented Kalman filter)融合的动力锂电池SOC估算方法。构建动力锂电池SOC估算窗口LSTM结构，通过动力电池电流、电压、温度并结合历史数据实时预测动力电池SOC训练网络；设计动力锂电池SOC估算UKF算法，提出融合策略。实验表明，研究窗口LSTM+UKF融合动力锂电池SOC估算方法RMSE、MAX、MAE分别为1.13%、1.74%、0.39%，相较于加窗LSTM网络提升了动力锂电池SOC估算的准确性、稳定性。

LSTM+UKF fusion-based SOC estimation method for powered lithium batteries
LI Hongpei, LIU Guixiong, DENG Wei
School of Mechanical and Automotive Engineering, South China University of Technology, Guangzhou 510640, China
Abstract: In order to improve the accuracy and stability of power battery state of charge SOC estimation, this paper proposes a power lithium battery SOC estimation method based on the fusion of LSTM + UKF (long short term memory+unscented Kalman filter). Firstly, the structure of power lithium battery SOC estimation window LSTM is constructed, and the power battery SOC training network is predicted in real time by combining the power battery current, voltage and temperature with the historical data. Secondly, the UKF Algorithm for SOC estimation of power lithium battery is designed, and the fusion strategy is proposed. The experimental results show that the research window LSTM + UKF fusion power lithium battery SOC estimation methods RMSE, MAX and MAE are 1.13%, 1.74% and 0.39% respectively, which improves the accuracy and stability of power lithium battery SOC estimation.
Keywords: power lithium battery;state of charge;long short term memory;unscented Kalman filter;algorithm fusion
2022, 48(8):22-28  收稿日期: 2022-06-07;收到修改稿日期: 2022-07-09
基金项目: 广东省重点领域研发计划项目 (2019B090908003)
作者简介: 李泓沛(1996-)，男，广东珠海市人，硕士研究生，专业方向为智能化检测与仪器研究
参考文献
[1] 石伟杰, 王海民. 基于锂离子电池热特性的SOH在线诊断模型研究[J]. 仪器仪表学报, 2020, 41: 206-216
[2] 夏克刚, 钱祥忠, 余懿衡. 联合精确估算电池的健康状态[J]. 电子测量技术, 2019, 42: 25-30
[3] 郝世宇, 殷会飞, 杨茹. 基于AEKF的锂离子动力电池荷电状态估计[J]. 国外电子测量技术, 2021, 40: 49-53
[4] ANTON J C A, NIETO P J G, VIEJO C B. Support vector machines used to estimate the battery state of charge[J]. IEEE Transactions on Power Electronics, 2013, 28: 5919-5926
[5] DU J N, LIU Z T, WANG Y Y. State of charge estimation for Li-ion battery based on model from extreme learning machine[J]. Control Engineering Practice, 2014, 26: 11-19
[6] 赵钢, 孙豪赛, 罗淑贞. 基于BP神经网络的动力电池SOC估算[J]. 电源技术, 2016, 40: 818-819
[7] CHAOUI H, IBE-EKEOCHA C C. State of charge and state of health estimation for lithium batteries using recurrent neural networks[J]. IEEE Transactions on Vehicular Technology, 2017, 66: 8773-8783
[8] CHEMALI E, KOLLMEYER P J, PREINDL M. Long short-term memory networks for accurate state-of-charge estimation of Li-ion batteries[J]. IEEE Transactions on Industrial Electronics, 2018, 65: 6730-6739
[9] SONG X, YANG F, WANG D. Combined CNN-LSTM network for state-of-charge estimation of lithium-ion batteries[J]. IEEE Access, 2019, 7: 88894-88902
[10] TIAN Y, LAI R, LI X. A combined method for state-of-charge estimation for lithium-ion batteries using a long short-term memory network and an adaptive cubature Kalman filter[J]. Applied Energy, 2020: 265
[11] YANG F, ZHANG S, LI W. State-of-charge estimation of lithium-ion batteries using LSTM and UKF[J]. Energy, 2020: 201
[12] 耿攀, 许梦华, 薛士龙. 基于LSTM循环神经网络的电池SOC预测方法[J]. 上海海事大学学报, 2019, 40: 120-126
[13] 李超然, 肖飞, 樊亚翔. 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: 2051-2062
[14] 李泓沛, 刘桂雄. 加窗LSTM网络的动力电池SOC预测优化算法[J]. 中国测试, 2021, 47: 87-91
[15] YU Y, SI X S, HU C H. A review of recurrent neural networks: LSTM cells and network architectures[J]. Neural Computation, 2019, 31: 1235-1270
[16] 刘桂雄, 李泓沛, 李远茂. 一种模拟车辆工况动力电池荷电状态SOC估算优化方法: CN111426957B[P]. 2021-06-11.
[17] 刘桂雄, 李远茂, 李泓沛. 一种LSTM网络辅助卡尔曼滤波SOC估算方法: CN113433457A[P]. 2021-09-24.
[18] US ABC. Electric vehicle battery test procedures manual revision 2[S], 1996.