A Data-Driven Thermal Runaway Warning Method for Lithium-Ion Batteries Under Mechanical Abuse

Mechanical abuse is a cause of thermal runaway (TR) for lithium-ion batteries (LIBs). Developing an effective method to predict TR for LIBs under mechanical abuse is crucial for improving safety, which can enable drivers to get out of the car before the TR. However, TR under different mechanical conditions is complex and uncertain, and TR behavior prediction of LIBs coupled with mechanical–electrical–thermal is challenging. Therefore, an experimental platform was designed for conducting a series of mechanical abuse experiments for LIBs, and a mechanical–electrical–thermal coupling model for LIBs was established to supplement the experimental data for model training. Then, a data-driven model named TR temperature prediction neural network (TRTPNN) was constructed to decompose the battery prediction task into normal and failure states by setting a model-switching strategy, which can extract deeper information about the characteristic variables in different stages. The proposed model reduces an average mean absolute error (MAE) and mean absolute percentage error (MAPE) in different experimental conditions compared to the single neural network model by 17.6% and 41.7%, respectively. Finally, a multistage TR warning strategy based on the TRTPNN model is presented, triggering three alarm levels before the TR in all tested batteries.

• Record URL:
  • https://doi.org/10.1109/TTE.2024.3459038
• Availability:
  • Find a library where document is available. Order URL: http://worldcat.org/issn/23327782
• Supplemental Notes:
  • Copyright © 2025, IEEE.
• Authors:
  • Li, Jie
  • Zhang, Yunlong
  • Yuan, Boxing
  • He, Yongquan
  • Zhu, Wei
  • 0000-0001-7878-9695
• Publication Date: 2025-4

Language

• English

Media Info

• Media Type: Web
• Features: References;
• Pagination: pp 5169-5179
• Serial:
  • IEEE Transactions on Transportation Electrification
  • Volume: 11
  • Issue Number: 2
  • Publisher: Institute of Electrical and Electronics Engineers (IEEE)
  • ISSN: 2332-7782
  • Serial URL: http://ieeexplore.ieee.org/servlet/opac?punumber=6687316

Subject/Index Terms

• TRT Terms: Lithium batteries; Neural networks; Predictive models; Temperature; Vehicle safety; Warning systems
• Subject Areas: Energy; Highways; Maintenance and Preservation; Safety and Human Factors; Vehicles and Equipment;

Filing Info

• Accession Number: 01958200
• Record Type: Publication
• Files: TRIS
• Created Date: Jun 18 2025 2:41PM