Search

Early LSTM-Based Early Warning System for Thermal Runaway of Lithium-Ion Batteries Using an ESP32-S3 Edge Computing Platform

Author(s)

Haojie Ma, Yichi Zhang, Junhao Liang, Zhichao Zhang*

Affiliation(s)

North China University of Science and Technology, Tangshan, China *Corresponding Author

Abstract

An early warning system for lithium battery thermal runaway during charging, designed as an independent external device for new energy vehicles, electric bicycles, and portable energy storage devices. The system integrates DS18B20 temperature sensors and three-in-one gas concentration sensors (CH₄, CO, C₂H₄) to capture multi-source data. An LSTM (Long Short-Term Memory) time-series prediction model is deployed on an ESP32-S3 microcontroller using edge computing, enabling real-time trend prediction of temperature and characteristic gas concentrations several minutes in advance. Compared to traditional threshold-based Battery Management Systems (BMS), this approach shifts from post-accident alarms to pre-accident prevention, significantly improving early detection accuracy and reducing false alarms. The device features independent power supply, wireless communication (Wi-Fi/Bluetooth), and a mobile app for real-time monitoring and alerts. With low cost (basic version under 100 RMB), easy installation, and broad compatibility, the product fills a market gap for consumer-grade thermal runaway warning solutions. Experimental results show millisecond-level inference latency and reliable early warning capability, offering strong potential for large-scale adoption and enhanced lithium battery charging safety.

Keywords

Lithium Battery; LSTM; ESP32

References

[1] Fernandes Y, Bry A, Persis D S. Identification and quantification of gases emitted during abuse tests by overcharge of a commercial Li-ion battery[J]. Journal of Power Sources,2018,389. [2] P.W. Nebiker, R.E. Pleisch. Photoacoustic gas detection for fire warning[J]. Fire Safety Journal,2001,36(2). [3] Ohsaki T, Kishi T, Kuboki T, et al. Overcharge reaction of lithium-ion batteries[J]. Journal of Power Sources, 2005, 146(1): 97-100. [4] Chunli Wang, Limiao Gong, Ping Kang, Yechao Tan, Mingming Li. Research on early warning system for lithium-ion battery energy storage power station[J]. Energy Storage Science and Technology, 2018,7(06):1152-1158. [5] Mingmin Wang, Lei Sun,Pengyu Guo, Dongliang Guo, Lantian Zhao, Yang Jin. Overcharge thermal runaway characteristics of lithium iron phosphate energy storage battery module based on online gas monitoring[J]. High Voltage Engineering, 2021, 47(01): 279-286.