Experimental Study on Thermal Runaway Characteristics and Gas Generation Behavior Induced by Overcharge: Single Cells and Modules of Large-Capacity Ncm Lithium-Ion Battery

Abstract

Large-capacity lithium nickel cobalt manganese oxide (NCM) batteries are widely used in electric vehicles (EVs) due to their high energy density. With its widespread applications and continuously developing charging technology, overcharge-induced fire incidents have been frequently reported globally. In this work, the thermal runaway (TR) characteristics and gas generation behaviors of 95 Ah NCM523 battery cells and modules induced by overcharge were investigated under different charging rates. Experimental measurements included video recording, surface temperature, heat release rate, mass loss, and evolved gases. Violent flame ejection was observed during TR with increasing overcharging rates, accompanied by a reduction in the angle between the ejected flames from the safety valve and ruptured sites. With an increase in the overcharging rate, a decrease in the proportion of heat generated by side reactions was identified. Much higher heat generation ratios from side reactions were recorded for battery modules compared to single cells. For single cells, higher concentrations of H2, CO, and CO2 were observed at high overcharging rates. In all tests H2 was released consistently earlier than CO and CO2. Based on experimental results and observations, a safety assessment scoring system was established for evaluating TR risk and hazard. It was found that both increased with the overcharging rate, with battery modules consistently having higher TR risks than single cells. The present findings provide not only the mechanisms governing TR and gas generation of large capacity NCM lithium-ion batteries but critical insights into charging optimization which are valuable for the EV industry to develop and assess the safety of charging systems for these types of batteries.

Keywords: NCM battery, Overcharge, Thermal runaway, Safety assessment, Charging optimization