Improving Lithium-ion Battery Safety with Early Thermal Runaway Diagnosis via EIS and Gas sensors

The increase of Electric Vehicles (EV) in the market, and therefore, the increase of isolated Thermal Runaway events has highlighted the need for measures to be taken in this matter, in response to the safety concerns among the industry and society. A Thermal Runaway event can happen over a certain period of time due to electrical or thermal abuse, or can happen suddenly due to mechanical abuse (nail penetration or crush). Regulations such as UN ECE R100 or GTR-EVS have added safety norms by stating that a Thermal Runaway event must be detected 5 minutes before it occurs. Thus, a redundant and reliable early fault detection method needs to be stablished. A Thermal Runaway event can be monitored by conventional sensors inside a BMS (voltage, current or temperature). Nevertheless, the effect of the Thermal Runaway event on these parameters becomes observable when the process is already underway, leaving little time for reaction. The Electrochemical Impedance Spectroscopy (EIS) is dependent on cell internal temperature, and it is stable in almost the whole SoC range [1], so cell internal temperature at a Thermal Runaway event is observed before it is measurable on the cell surface by thermocouples, providing and early warning of the event [2]. Furthermore, the measurement of first venting gases from the cell at the Thermal Runaway event allows a significantly faster detection of the event than conventional BMS sensors and can even detect a sudden Thermal Runaway event, adding robustness to the system. In this work, we present a redundant Thermal Runaway detection method based on EIS and Gas Sensors. In this regard, overcharge and overtemperature abuse tests are carried out to a NMC 21700 cylindrical cell, measuring CO₂ concentration, H₂ concentration, air pressure, temperature and humidity with a state-of-the-art sensor, alongside a portable EIS device. The results will be analysed in order to prove that a Thermal Runaway event can be detected by means of EIS measurements and that the Gas Sensors also detects the event successfully. This enables the possibility to give a warning signal prior to the 5 minutes window specified in the standards.