Thermal Runaway of Li-ion batteries is a major safety risk, and the propagation of fire from one cell to the neighboring cells is a huge challenge. Furthermore, the risk of damage to loose cells in the waste and recycling stream can result in catastrophic fires leading to property damage and injury. During storage and transport, batteries from collection centers awaiting recycling are typically housed in metal 55 gallon (0.208 m3) drums. Several accidents resulting in thermal runaway have occurred, without being observed from the outside until gas pressure caused the vessel to rupture. Early and robust detection for thermal runaway events could enable protection and mitigation to reduce the risks. Previous studies have used voltage or current to detect possible thermal runaway events, assuming the cells are assembled into a pack where electrical and thermal measurements are feasible. However, in this case, without any battery management system that measures the current, voltage, and temperature of the individual cells, the only information we can get is from a few sensors at the lid of the drum. To address the challenges of storage and transport, a Li-ion battery thermal runaway early detection method based on gas sensing is developed. To compare the performance of several detection methods, a 0.208 m3 cylindrical storage drum filled with fully charged 18650 cells, is simulated in COMSOL. In the simulation, an 18650 NMC cell, located in the center of the drum, as shown in Fig. 1, triggers a hard internal short circuit and leads to thermal runaway. A single cell thermal runaway can heat neighboring cells and triggers new thermal runaway events, thus causing a chain reaction of more and more thermal runaway cells [1]. A nine-cell system, with three cells in parallel and three cells in series, is simulated. Each battery is spaced 2 mm from each other, and the free space is assumed to be filled with air. The results for temperature evolution of each cell during thermal runaway propagation is shown in Fig. 2. As seen from the figure, the second cell triggers thermal runaway after 710 seconds. The remaining cells quickly enter the thermal runaway state after the second cell. The detection of thermal runaway must be made before this time if corrective action is to be applied. So, this time is used as the benchmark of thermal runaway detection in this study, defined as propagation critical time. To detect possible thermal runaway, temperature sensors are located on the surface of the drum as shown in Fig. 1. From the simulation result, at 710 seconds, the maximum surface temperature change on the drum surface is less than 0.001 °C. In this simulation, conventional surface temperature sensing approach (e.g., Infrared Imaging) cannot detect the thermal runaway event before the propagation critical time of 710s when the second cell starts a thermal runaway event. Extending the model of [2] inside a drum full of 18650 cells, the detection of a single thermal runaway event by CO2 gas sensors is investigated. Li-ion battery vents significant amount of gas during thermal runaway, and due to SEI decomposition reaction at the early stage of thermal runaway [3], CO2 is generated and vented from battery to the drum space. This gas sensing approach then monitors the concentrations of CO2 inside the drum, and cell venting during thermal runaway can be detected when the concentrations of CO2 are high. The CO2 sensor is located at the center of the drum top surface in Fig .1, and the detection threshold of CO2 concentrations for a thermal runaway event is set to be 2000 ppm. Fig. 3 shows the evolution of CO2 concentrations with time. The initial gas venting speed will accelerate the CO2 species transport process, and CO2 propagates to most of the drum space in a short period. The detection can be made at 85 seconds, which is ahead of the propagation critical time 710 seconds. This study shows the faster response of gas sensing comparing to drum surface temperature sensing in a cylindrical drum. With the use of gas sensing approach in large battery storage facilities, the speed of thermal runaway event detection can be improved and precede the propagation of thermal runaway to additional cells. Reference [1] Kim, G-H.; Pesaran, A.; Smith, K. No. NREL/PR-540-43186. National Renewable Energy Lab.(NREL), Golden, CO (United States), 2008. [2] Cai, T.; Stefanopoulou, A.; Siegel, J. In Dynamic Systems and Control Conference, ASME: 2018. [3] Feng, X.; Ouyang, M.; Liu, X.; Lu, L.; Xia, Y.; He, X. Energy Storage Materials 2018, 10, 246-267. Figure 1
Read full abstract
Jul 1, 2020
R Rugsaj + 1
Open Access
