journal article·2026-10

Transformer-based autoencoder for robust detection of battery anomalies using temperature sensors

Tymon B. Nieduzak, Hanita K. Haller, Dhananjay Narayanan, Maria Q. Feng

Publisher page

Abstract

Thermal runaway (TR) in lithium-ion batteries (LIBs) poses a critical safety risk in electric vehicles (EVs), where rapid failure can lead to fire or explosion. Early detection remains challenging because varying environmental conditions, operating profiles, and runaway pathways produce diverse thermal signatures that conventional temperature threshold-based methods often fail to capture. This work presents a machine-learning framework for early detection of LIB thermal abnormalities using only temperature sensors available in existing battery management systems. A transformer-based autoencoder is trained on large datasets of normal LIB cycling across multiple chemistries, cell formats, and operating conditions to learn operational thermal dynamics. The model reconstructs recent temperature histories, and deviations from normal behavior are identified through increased reconstruction error rather than absolute temperature magnitude. At a reconstruction error threshold of 200, the system achieves a 98.9% detection rate for diverse mechanically induced thermal anomalies with a 3.7% false positive rate and detects 100% of TR events. For runaway cases, the framework provides an average detection lead time of 58.3 seconds relative to a 100°C threshold, with lead times ranging up to 15.1 minutes. Inference latency is <0.054 seconds per evaluation window, enabling real-time deployment and providing a scalable pathway for next-generation battery monitoring.
Source: openaire · Quality: candidate unverified · Source record ↗

Research context

OpenAlex topics: Advanced Battery Technologies Research · Advanced Chemical Sensor Technologies · Anomaly Detection Techniques and Applications

SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.

References (41)

Show source-provided references
1. Li-ion battery materials: present and future10.1016/j.mattod.2014.10.040 ↗
2. Ageing mechanisms in lithium-ion batteries10.1016/j.jpowsour.2005.01.006 ↗
3. Potential of lithium-ion batteries in renewable energy10.1016/j.renene.2014.11.058 ↗
4. Li‐ion batteries: basics, progress, and challenges10.1002/ese3.95 ↗
5. Research progress on thermal runaway warning methods and fire extinguishing technologies for lithium-ion batteries10.3390/wevj16020081 ↗
6. Experimental study of internal and external short circuits of commercial automotive pouch lithium-ion cells10.1016/j.est.2018.01.015 ↗
7. Characterization of penetration induced thermal runaway propagation process within a large format lithium ion battery module10.1016/j.jpowsour.2014.11.017 ↗
8. Mechanical properties and thermal runaway study of automotive lithium-ion power batteries
9. Electrical behavior of overdischarge-induced internal short circuit in lithium-ion cells10.1016/j.electacta.2018.05.048 ↗
10. Investigation of a commercial lithium-ion battery under overcharge/over-discharge failure conditions10.1039/C8RA05564E ↗
11. A mechanism-data driven resistance transfer algorithm for lithium-ion batteries and its application to thermal modeling10.1016/j.est.2024.114066 ↗
12. A brief review on key technologies in the battery management system of electric vehicles10.1007/s11465-018-0516-8 ↗
13. Why traditional BMS Can't detect thermal runaway fast enough
14. On-Board implementation of thermal runaway detection in lithium-ion battery packs: methods, metrics, and challenges10.3390/en19030858 ↗
15. Research on thermal runaway warning threshold and machine learning model for large capacity energy storage batteries based on electric-thermal-mechanical-gas coupling10.1016/j.psep.2026.108471 ↗
16. Failure modes and mechanisms for rechargeable Lithium-based batteries: a state-of-the-art review10.1007/s00707-018-2327-8 ↗
17. Cell-to-cell variability in Li-ion battery thermal runaway: experimental testing, statistical analysis, and kinetic modeling10.1016/j.est.2022.106024 ↗
18. A critical review of thermal runaway prediction and early-warning methods for lithium-ion batteries10.34133/energymatadv.0008 ↗
19. A novel battery abnormality detection method using interpretable autoencoder10.1016/j.apenergy.2022.120312 ↗
20. A transformer guided multi modal learning framework for predictive and causal assessment of thermal runaway in high energy batteries10.1038/s41598-025-20886-x ↗
21. Characteristics of battery management systems of electric vehicles with consideration of the active and passive cell balancing process10.3390/wevj12030120 ↗
22. Battery management systems—Challenges and some solutions10.3390/en13112825 ↗
23. Evaluation of detection criteria for thermal runaway experiments on commercial cells for electric vehicles10.1016/j.egyr.2023.07.001 ↗
24. Anomaly detection during lithium-ion battery qualification testing
25. Hybrid machine learning framework for predictive maintenance and anomaly detection in lithium-ion batteries using enhanced random forest10.1038/s41598-025-90810-w ↗
26. Pang
27. Battery prognostics and health management from a machine learning perspective10.1016/j.jpowsour.2023.233474 ↗
28. Battery thermal runaway fault prognosis in electric vehicles based on abnormal heat generation and deep learning algorithms10.1109/TPEL.2022.3150026 ↗
29. Survey of lithium-ion battery Anomaly detection methods in electric vehicles
30. State-of-health estimation with anomalous aging indicator detection of lithium-ion batteries using regression generative adversarial network10.1109/TIE.2022.3170630 ↗
31. Realistic fault detection of li-ion battery via dynamical deep learning10.1038/s41467-023-41226-5 ↗
32. Deep learning‐driven modeling for thermal runaway warning during lithium‐ion battery charging in electric vehicles10.1002/adts.202501438 ↗
33. Battery fault detection and early warning for electric vehicles: a deep learning-powered end-to-end solution
34. BatteryArchive.org, (n.d.). https://www.batteryarchive.org/(accessed January 19, 2026).
35. Howey · 2017
36. Degradation of commercial lithium-ion cells as a function of chemistry and cycling conditions10.1149/1945-7111/abae37 ↗
37. Degradation-safety analytics in lithium-ion cells: part I. Aging under charge/discharge cycling10.1149/1945-7111/abc8c0 ↗
38. Mechanically induced thermal runaway severity analysis for Li-ion batteries10.1016/j.est.2023.106798 ↗
39. Mitigating lithium-ion cell thermal runaway via selective trace H2 sensing10.1016/j.xcrp.2025.102859 ↗
40. STM32H7 - arm Cortex-M7 and Cortex-M4 MCUs (480 MHz) - STMicroelectronics, (n.d.). https://www.st.com/en/microcontrollers-microprocessors/stm32h7-series.html (accessed April 21, 2026).
41. S32G vehicle network processors, (n.d.). https://www.nxp.com/products/processors-and-microcontrollers/s32-automotive-platform/s32g-vehicle-network-processors:S32G-PROCESSORS (accessed April 21, 2026).