LI-ion NOSE : A Multi-Parametric Sensor Fusion Framework for Pre-Thermal Runaway Detection

Lithium-ion batteries are the dominant power source for modern electronics, yet they remain vulnerable to “thermal runaway”—a catastrophic failure mode often preceded by electrolyte off-gassing 5–20 minutes before critical temperature spikes occur. Addressing the latency limitations of traditional temperature-only Battery Management Systems (BMS), this paper presents a low- cost, standalone early warning system based on the ESP8266 microcontroller. The proposed prototype utilizes a Sensor Fusion algorithm that correlates real- time gas anomalies (using an MQ-135 sensor) with thermal Rate-of-Rise (RoR) analysis (using a DS18B20 probe). By implementing a tiered “Safety Ladder” logic, where gas detection triggers a pre-warning and rapid heating triggers a critical alarm, the system significantly reduces false positives while capitalizing on the “Golden Window” of pre-failure detection. Experimental validation using combustion surrogates demonstrates that this multi-modal approach offers a reaction time advantage of several minutes compared to conventional thermal monitoring.