TY - JOUR
T1 - A deep learning method for online capacity estimation of lithium-ion batteries
AU - Shen, S.
AU - Sadoughi, Mohammadkazem
AU - Chen, Xiangyi
AU - Hong, Mingyi
AU - Hu, Chao
N1 - Publisher Copyright:
© 2019 Elsevier Ltd
PY - 2019/10
Y1 - 2019/10
N2 - The past two decades have seen an increasing usage of lithium-ion (Li-ion) rechargeable batteries in diverse applications including consumer electronics, power backup, and grid-scale energy storage. To guarantee safe and reliable operation of a Li-ion battery pack, battery management systems (BMSs) should possess the capability to monitor, in real time, the state of health (SOH) of the individual cells in the pack. This paper presents a deep learning method which utilizes deep convolutional neural network (DCNN) for cell-level capacity estimation based on the voltage, current, and charge capacity measurements during a partial charge cycle. The unique features of DCNN include the local connectivity and shared weights, which enable the model to accurately estimate battery capacity using the measurements during charge. To the best of our knowledge, this is one of the first attempts to apply deep learning to the online capacity estimation of Li-ion batteries. Ten-year daily cycling data from eight implantable Li-ion cells and half-year cycling data from 20 18650 Li-ion cells were utilized to verify the performance of the proposed deep learning method. Compared with traditional machine learning methods such as shallow neural networks and relevance vector machine (RVM), the proposed deep learning method is demonstrated to produce higher accuracy and robustness in the online estimation of Li-ion battery capacity.
AB - The past two decades have seen an increasing usage of lithium-ion (Li-ion) rechargeable batteries in diverse applications including consumer electronics, power backup, and grid-scale energy storage. To guarantee safe and reliable operation of a Li-ion battery pack, battery management systems (BMSs) should possess the capability to monitor, in real time, the state of health (SOH) of the individual cells in the pack. This paper presents a deep learning method which utilizes deep convolutional neural network (DCNN) for cell-level capacity estimation based on the voltage, current, and charge capacity measurements during a partial charge cycle. The unique features of DCNN include the local connectivity and shared weights, which enable the model to accurately estimate battery capacity using the measurements during charge. To the best of our knowledge, this is one of the first attempts to apply deep learning to the online capacity estimation of Li-ion batteries. Ten-year daily cycling data from eight implantable Li-ion cells and half-year cycling data from 20 18650 Li-ion cells were utilized to verify the performance of the proposed deep learning method. Compared with traditional machine learning methods such as shallow neural networks and relevance vector machine (RVM), the proposed deep learning method is demonstrated to produce higher accuracy and robustness in the online estimation of Li-ion battery capacity.
KW - Capacity estimation
KW - Deep learning
KW - Health monitoring
KW - Lithium-ion batteries
UR - http://www.scopus.com/inward/record.url?scp=85070187168&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85070187168&partnerID=8YFLogxK
U2 - 10.1016/j.est.2019.100817
DO - 10.1016/j.est.2019.100817
M3 - Article
AN - SCOPUS:85070187168
SN - 2352-152X
VL - 25
JO - Journal of Energy Storage
JF - Journal of Energy Storage
M1 - 100817
ER -