A Novel Voltage-Abnormal Cell Detection Method for Lithium-Ion Battery Mass Production Based on Data-Driven Model with Multi-Source Time Series Data

Xiang Wang, Jianjun He (), Fuxin Huang, Zhenjie Liu, Aibin Deng and Rihui Long
Additional contact information
Xiang Wang: The School of Automation, Central South University, Changsha 410038, China
Jianjun He: The School of Automation, Central South University, Changsha 410038, China
Fuxin Huang: The School of Automation, Central South University, Changsha 410038, China
Zhenjie Liu: The School of Automation, Central South University, Changsha 410038, China
Aibin Deng: The School of Automation, Central South University, Changsha 410038, China
Rihui Long: The School of Automation, Central South University, Changsha 410038, China

Energies, 2024, vol. 17, issue 14, 1-14

Abstract: Before leaving the factory, lithium-ion battery (LIB) cells are screened to exclude voltage-abnormal cells, which can increase the fault rate, troubleshooting difficulty, and degrade pack performance. However, the time interval to obtain the detection results through the existing voltage-abnormal cell method is too long, which can seriously affect production efficiency and delay shipment, especially in the mass production of LIBs when facing a large number of time-critical orders. In this paper, we propose a data-driven voltage-abnormal cell detection method, using a fast model with simple architecture, which can detect voltage-abnormal cells based on the multi-source time series data of the LIB without a time interval. Firstly, our method transforms the different source data of a cell into a multi-source time series data representation and utilizes a recurrent-based data embedding to model the relation within it. Then, a simplified MobileNet is used to extract hidden feature from the embedded data. Finally, we detect the voltage-abnormal cells according to the hidden feature with a cell classification head. The experiment results show that the accuracy and average running time of our model on the voltage-abnormal cell detection task is 95.42% and 0.0509 ms per sample, which is a considerable improvement over existing methods.

Keywords: lithium-ion battery production; data-driven model; anomaly detection; multi-source time series data (search for similar items in EconPapers)
JEL-codes: Q Q0 Q4 Q40 Q41 Q42 Q43 Q47 Q48 Q49 (search for similar items in EconPapers)
Date: 2024
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://www.mdpi.com/1996-1073/17/14/3472/pdf (application/pdf)
https://www.mdpi.com/1996-1073/17/14/3472/ (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:gam:jeners:v:17:y:2024:i:14:p:3472-:d:1435148

Access Statistics for this article

Energies is currently edited by Ms. Agatha Cao

More articles in Energies from MDPI
Bibliographic data for series maintained by MDPI Indexing Manager ().