Design of Thermal Runaway Early Warning and Safety Management System for Polymer Lithium-ion Batteries in Wearable Pet Devices

Abstract

For smart pet collars and health monitors equipped with thin polymer lithium batteries, bending, extrusion and wide temperature swings easily trigger latent thermal runaway, while traditional BMS only supports basic overcharge/over-discharge protection without early detection and graded intervention. This paper analyzes electro-mechanical-environment coupled thermal runaway mechanisms, builds multi-source time-series datasets, and proposes a physics-driven lightweight embedded warning model. A complete perception-edge computing graded safety hardware-software system is developed and validated by finite element simulation, calorimetry and prototype tests. The model reaches 96.4% accuracy for early thermal runaway, issuing alarms 128–176 s earlier than single temperature threshold schemes. The overall system power consumption is only 8.7 mW, enabling staged current limiting and power cut-off to curb thermal runaway evolution. This research provides theoretical and engineering support for battery safety design of pet wearable electronics.