Detecting Abnormality of Battery Lifetime from First-Cycle Data Using Few-Shot Learning

Xiaopeng TANG; Xin LAI; Changfu ZOU; Yuanqiang ZHOU; Jiajun ZHU; Yuejiu ZHENG; Furong GAO

doi:10.1002/advs.202305315

Detecting Abnormality of Battery Lifetime from First-Cycle Data Using Few-Shot Learning

Xiaopeng TANG, Xin LAI, Changfu ZOU^*, Yuanqiang ZHOU, Jiajun ZHU, Yuejiu ZHENG, Furong GAO

^*Corresponding author for this work

Science Unit

Research output: Journal Publications › Journal Article (refereed) › peer-review

17 Citations (Scopus)

Abstract

The service life of large battery packs can be significantly influenced by only one or two abnormal cells with faster aging rates. However, the early-stage identification of lifetime abnormality is challenging due to the low abnormal rate and imperceptible initial performance deviations. This work proposes a lifetime abnormality detection method for batteries based on few-shot learning and using only the first-cycle aging data. Verified with the largest known dataset with 215 commercial lithium-ion batteries, the method can identify all abnormal batteries, with a false alarm rate of only 3.8%. It is also found that any capacity and resistance-based approach can easily fail to screen out a large proportion of the abnormal batteries, which should be given enough attention. This work highlights the opportunities to diagnose lifetime abnormalities via “big data” analysis, without requiring additional experimental effort or battery sensors, thereby leading to extended battery life, increased cost-benefit, and improved environmental friendliness.

Original language	English
Article number	2305315
Journal	Advanced Science
Volume	11
Issue number	6
Early online date	11 Dec 2023
DOIs	https://doi.org/10.1002/advs.202305315
Publication status	Published - 9 Feb 2024

Bibliographical note

Publisher Copyright:
© 2023 The Authors. Advanced Science published by Wiley-VCH GmbH.

Funding

The first author would like to thank the continuing support from the Guangzhou HKUST Fok Ying Tung Research Institute during Hong Kong's unrest and the outbreak of Covid-19. The authors would like to thank Dr. Weihan Li from RWTH Aachen University, the first author of Ref. [26], for providing technical support when we tested our method on their dataset. We would also like to thank Ms Wenying Huang and Mr Jun Yuan from Far East Battery for their recommendations onaging experiment design when using batteries produced by their company. This work was supported, in part, by the Hong Kong RGC Postdoctoral Fellowship Scheme (PDFS2122-6S06), Hong Kong Research Grant Council under grant 16208520, the National Natural Science Foundation of China (51977131,52277223), the Natural Science Foundation of Shanghai (19ZR1435800), Guangdong Scientifi,c and Technological Project (2019A050516002), the Foshan-HKUST Projects Program (FSUST20-FYTRI12F), and the Swedish Energy Agency under the Vehicle Strategic Research and Innovation Program (Grant No. 50187-1).

Keywords

big data
early-stage detection
few-shot learning
lifetime abnormality
lithium-ion battery

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1002/advs.202305315Licence: CC BY

https://onlinelibrary.wiley.com/doi/abs/10.1002/advs.202305315

Cite this

@article{0fb0d0ab0caa49839a06a93111d959be,

title = "Detecting Abnormality of Battery Lifetime from First-Cycle Data Using Few-Shot Learning",

abstract = "The service life of large battery packs can be significantly influenced by only one or two abnormal cells with faster aging rates. However, the early-stage identification of lifetime abnormality is challenging due to the low abnormal rate and imperceptible initial performance deviations. This work proposes a lifetime abnormality detection method for batteries based on few-shot learning and using only the first-cycle aging data. Verified with the largest known dataset with 215 commercial lithium-ion batteries, the method can identify all abnormal batteries, with a false alarm rate of only 3.8%. It is also found that any capacity and resistance-based approach can easily fail to screen out a large proportion of the abnormal batteries, which should be given enough attention. This work highlights the opportunities to diagnose lifetime abnormalities via “big data” analysis, without requiring additional experimental effort or battery sensors, thereby leading to extended battery life, increased cost-benefit, and improved environmental friendliness.",

keywords = "big data, early-stage detection, few-shot learning, lifetime abnormality, lithium-ion battery",

author = "Xiaopeng TANG and Xin LAI and Changfu ZOU and Yuanqiang ZHOU and Jiajun ZHU and Yuejiu ZHENG and Furong GAO",

note = "Publisher Copyright: {\textcopyright} 2023 The Authors. Advanced Science published by Wiley-VCH GmbH.",

year = "2024",

month = feb,

day = "9",

doi = "10.1002/advs.202305315",

language = "English",

volume = "11",

journal = "Advanced Science",

issn = "2198-3844",

publisher = "Wiley - VCH Verlag GmbH & CO. KGaA",

number = "6",

}

TY - JOUR

T1 - Detecting Abnormality of Battery Lifetime from First-Cycle Data Using Few-Shot Learning

AU - TANG, Xiaopeng

AU - LAI, Xin

AU - ZOU, Changfu

AU - ZHOU, Yuanqiang

AU - ZHU, Jiajun

AU - ZHENG, Yuejiu

AU - GAO, Furong

PY - 2024/2/9

Y1 - 2024/2/9

N2 - The service life of large battery packs can be significantly influenced by only one or two abnormal cells with faster aging rates. However, the early-stage identification of lifetime abnormality is challenging due to the low abnormal rate and imperceptible initial performance deviations. This work proposes a lifetime abnormality detection method for batteries based on few-shot learning and using only the first-cycle aging data. Verified with the largest known dataset with 215 commercial lithium-ion batteries, the method can identify all abnormal batteries, with a false alarm rate of only 3.8%. It is also found that any capacity and resistance-based approach can easily fail to screen out a large proportion of the abnormal batteries, which should be given enough attention. This work highlights the opportunities to diagnose lifetime abnormalities via “big data” analysis, without requiring additional experimental effort or battery sensors, thereby leading to extended battery life, increased cost-benefit, and improved environmental friendliness.

AB - The service life of large battery packs can be significantly influenced by only one or two abnormal cells with faster aging rates. However, the early-stage identification of lifetime abnormality is challenging due to the low abnormal rate and imperceptible initial performance deviations. This work proposes a lifetime abnormality detection method for batteries based on few-shot learning and using only the first-cycle aging data. Verified with the largest known dataset with 215 commercial lithium-ion batteries, the method can identify all abnormal batteries, with a false alarm rate of only 3.8%. It is also found that any capacity and resistance-based approach can easily fail to screen out a large proportion of the abnormal batteries, which should be given enough attention. This work highlights the opportunities to diagnose lifetime abnormalities via “big data” analysis, without requiring additional experimental effort or battery sensors, thereby leading to extended battery life, increased cost-benefit, and improved environmental friendliness.

KW - big data

KW - early-stage detection

KW - few-shot learning

KW - lifetime abnormality

KW - lithium-ion battery

UR - http://www.scopus.com/inward/record.url?scp=85179362609&partnerID=8YFLogxK

U2 - 10.1002/advs.202305315

DO - 10.1002/advs.202305315

M3 - Journal Article (refereed)

C2 - 38081795

SN - 2198-3844

VL - 11

JO - Advanced Science

JF - Advanced Science

IS - 6

M1 - 2305315

ER -

Detecting Abnormality of Battery Lifetime from First-Cycle Data Using Few-Shot Learning

Abstract

Bibliographical note

Funding

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this