SOC estimation method of power battery based on LSTM-DaNN

doi:10.19799/j.cnki.2095-4239.2020.0200

Abstract

Abstract:

In this study, a long–short-term memory (LSTM) recurrent neural network is used to establish an estimation model of the state of charge (SOC) to estimate the SOC of power batteries. The model is trained and tested with a laboratory's constant current discharge data. The maximum absolute error is 2.7%. A further verification by the FSEC racing battery measured data shows a maximum test error of 3.9%. However, considering the complexity of the environment during the actual operation and the inconsistency caused by different driving habits to the power battery in engineering applications, training and testing must be performed according to the actual driving conditions of the vehicle. The SOC in driving conditions are directly translated from the battery management system (BMS) message; hence, we cannot affirm whether the SOC algorithm in the BMS is accurate. Consequently, the SOC in the driving conditions cannot be used as a label while training the model. At this time, the correct training label must be calculated or the existing model trained by the labeled data must be used. Its model parameters must then be dynamically adjusted based on the actual operating unlabeled data. This study takes the second method to solve the training problem of unlabeled data, proposing for the first time the combination of the domain adaptive neural network (DaNN) in transfer learning with LSTM to form the SOC estimation algorithm of LSTM-DaNN using the labeled data to train the LSTM model in advance, transfer its model parameters to LSTM-DaNN, and finally train the LSTM-DaNN model by combining the labeled and unlabeled data. The result shows that LSTM-DaNN can complete the training without the label of the actual driving condition (i.e., SOC) and ensure an absolute error of 4.8%. Compared with the model before the adaptive adjustment, the error decreases by 14.1%, and the absolute error is guaranteed to be <5%, meeting actual needs.

Key words: state of charge(SOC), long short-term memory(LSTM), transfer learning, domain adaptive neural network(DaNN)

CLC Number:

TM 912.9

Yiquan WANG, Bixiong HUANG, Xiao YAN, Xintian LIU, Ying WANG, Shuangyu LIU, Huayuan XU. SOC estimation method of power battery based on LSTM-DaNN[J]. Energy Storage Science and Technology, 2020, 9(6): 1969-1975.

Figures/Tables 12

Fig.1

Fig.2

Fig.3

Fig.4

Fig.5

Fig.6

Fig.7

Fig.8

Fig.9

Fig.10

Fig.11

Fig.12

References 13

1	孙巍. 国内近几年锂电池SOC估计方法综述[J]. 福建质量管理, 2019(3): 248.
	SUN Wei. Overview of SOC estimation methods for lithium batteries in China in recent years[J]. Fujian Quality Management, 2019(3): 248.
2	陈元丽,赵振东,陈素娟,等. 动力锂电池SOC估算方法综述[J]. 汽车科技, 2019(5): 65-69.
	CHEN Yuanli, ZHAO Zhendong, CHEN Sujuan, et al. Overview of SOC estimation methods for power lithium batteries[J]. Automotive Technology, 2019(5): 65-69.
3	胡耘. 动力电池荷电状态(SOC)估算方法综述[J]. 汽车实用技术, 2019(8): 36-38.
	HU Yun. Overview of estimation method of power battery state of charge(SOC)[J]. Automotive Practical Technology, 2019(8): 36-38.
4	沈佳妮, 贺益君, 马紫峰. 基于模型的锂离子电池SOC及SOH估计方法研究进展[J]. 化工学报, 2018, 69(1): 317-324.
	SHEN Jiani, HE Yijun, MA Zifeng. Research progress of model-based estimation methods for SOC and SOH of lithium ion batteries[J]. Journal of Chemical Engineering, 2018, 69(1): 317-324.
5	孙立珍. 锂离子电池SOC估计方法研究[D]. 呼和浩特: 内蒙古工业大学, 2019.
	SUN Lizhen. Research on SOC estimation method of lithium ion battery[D]. Huhhot: Inner Mongolia University of Technology, 2019.
6	周韦润, 姜文刚. 基于遗传算法优化扩展卡尔曼滤波的锂电池SOC估计[J].重庆理工大学学报(自然科学), 2019, 33(9): 33-39.
	ZHOU Weirun, JIANG Wengang. Lithium battery SOC estimation based on genetic algorithm optimized extended Kalman Filter[J]. Journal of Chongqing University of Technology(Natural Science), 2019, 33(9): 33-39.
7	Ma Y, Li X, Li G, et al. SOC oriented electrochemical-thermal coupled modeling for lithium-ion battery[J]. IEEE Access, 2019,7: 156136-156149.
8	FANG Linlin, LI Junqiu, PENG Bo. Online estimation and error analysis of both SOC and SOH of lithium-ion battery based on DEKF methond [J]. Energy Procedia, 2019, 158: 3008-3013.
9	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(9): 1453-1457.
	YAO Fang, ZHANG Nan, HUANG Kai. BP neural network algorithm based on LM for estimating SOC of lithium battery[J]. Chinese Journal of Power Sources, 2019(9): 1453-1457.
10	杨云龙. 基于改进型神经网络动力电池SOC估计研究[D]. 成都: 电子科技大学, 2019.
	YANG Yunlong. Research on SOC estimation of power battery based on improved neural network[D]. Chengdu: University of Electronic Science and Technology of China, 2019.
11	耿攀. 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(3): 120-126.
	GENG Pan. Battery SOC prediction method based on LSTM recurrent neural network[J]. Journal of Shanghai Maritime University, 2019, 40(3): 120-126.
12	PAN S J, YANG Q. A survey on transfer learning[J]. IEEE Tkde, 2010, 22(10): 1345-1359.
13	GHIFARY M, KLEIJN W B, ZHANG M. Domain adaptive neural networks for object recognition[C]//Pacific Rim International Conference on Artificial Intelligence. Springer, Cham, 2014: 898-904.

[1]	LI Na, BAI Kai, LIU Yu, WANG Kairang, GONG Yu, DONG Jianming. The energy storage system output control strategy to improve the short term wind power forecasting accuracy rate [J]. Energy Storage Science and Technology, 2018, 7(1): 100-.
[2]	LI Junhui, GAO Tianyu, ZHAO Bing, YAN Gangui, JIAO Jian. Inhibition of wind power fluctuations of battery energy storage system adaptive control strategy design [J]. Energy Storage Science and Technology, 2015, 4(3): 278-283.