📄 Mapping heat flow in prismatic battery modules during thermal runaway propagation using empirical data
Batteries & Supercaps, 2026,
Read more ↗
Yang Yang et al.
Abstract: To advance the electrification of the transport sector beyond passenger cars, electrifying heavy-duty trucks is essential. These vehicles typically use prismatic lithium-ion cells arranged in modules, separated by heat-insulating thermal pads that enhance safety during thermal runaway (TR). In this study, we developed and applied a method to map heat flow through various paths during TR propagation across three test cases with different thermal pads. The results were quantitatively evaluated using Sankey diagrams, a novel approach in this context. Using this method, we measured in situ thermal conductivity and found significant differences from standard reference values. As expected, lower in situ thermal conductivity increased the delay in thermal propagation. However, the method revealed that while the thermal pad remains the primary heat flow path during TR propagation, other contributors become significant if the pad has sufficiently low thermal conductivity. This finding is noteworthy, as the pad with the lowest conductivity nearly stops the propagation altogether, and attention to the other paths could be the key to achieving a full stop. We conclude that by investigating thermal pads under operational conditions, this study provides valuable insights into critical heat transfer paths and failure mechanisms, offering guidance on optimizing battery safety and lifespan.
📄 Modeling the interplay between aging and thermal runaway propagation in large-format lithium-ion batteries
Journal of Power Sources Advances, 2026,
Read more ↗
Yang Yang et al.
Abstract: Thermal runaway (TR) and its propagation (TRP) pose critical risks in the application of large-format lithium-ion batteries in heavy-duty electric vehicles. In this work, we apply a computational approach using a lumped heat release model. This model is calibrated with experimental data from accelerating rate calorimetry (ARC) and TRP tests to investigate battery aging effects on TR and TRP. It is seen that the simulations can effectively reproduce key experimental observations, such as TR onset temperature, maximum temperature, and TRP time. Furthermore, the influence of battery aging on TR behavior is investigated, specifically solid–electrolyte interphase (SEI) growth and electrolyte degradation. The findings reveal that aging significantly accelerates TR onset while lowering the heat release of batteries. The interplay between accelerated SEI layer growth and electrolyte degradation significantly influences TRP dynamics. Compared to new batteries, the total TRP time initially decreases during early aging, reaching 78% of the original TRP time at around 80% state of health (SOH). During late aging, TRP time slightly increases to 85% of the original time at 50% SOH. This computational approach provides crucial insights into the dynamic safety of aged batteries with regard to different combinations of electrolyte degradation and SEI thickness growth rate.
📕 Thermal Runaway in large-format lithium-ion batteries: Experimental, diagnostic, and modeling approaches for safer battery design
Acta Universitatis Upsaliensis (AUU), 2025,
Read more ↗
Yang Yang
Abstract: Ensuring the safety of lithium-ion batteries requires robust methods to study thermal runaway (TR) and its propagation (TRP). While accelerating rate calorimetry (ARC) has been the standard method, it is costly and limited in applicable cell sizes. This thesis develops empirical and novel approaches that provide cost-effective and scalable alternatives. First, TRP tests on 157 Ah LiNi0.8Mn0.1Co0.1O2 cells using widely available thermocouples were analyzed, enabling the estimation of onset and maximum temperatures, heat release, and temperature increase rates. Results showed close agreement with ARC, while offering broader applicability and lower complexity. Next, pouch and prismatic LiNi0.5Mn0.3Co0.2O2 cells were investigated with multidimensional sensors (i.e., force, gas, voltage, temperature), which allowed for a comprehensive safety characterization and revealed a consistent failure sequence of swelling, venting, gas emission, internal short circuit, and TR. While no significant format differences were found under overcharging, prismatic cells exhibited superior safety under overheating due to their higher mechanical strength and thermal dissipation. Scaling effects were then explored by comparing lab-scale coin cells (8.6 mAh) with industrial-scale cells up to 157 Ah, showing that small-scale tests are highly sensitive to the trigger methods, whereas industrial-scale cells yielded comparatively consistent normalized heat release, highlighting the limitations of downscaling. The TRP methodology was extended to map heat transfer in modules, where busbars and thermal pads were identified as critical heat conduction pathways, and in-situ measurements showed that thermal conductivity of pads under TR conditions deviated substantially from nominal values, strongly influencing TRP time. Finally, computational modeling was employed to simulate aging effects on TR, demonstrating that early aging accelerates TRP due to SEI growth, while late aging reduces total heat release due to further degradations but still sustains faster propagation than fresh batteries. Collectively, these studies integrate empirical diagnostics, module-level analysis, and computational modeling to provide a comprehensive picture of TR across scales, formats, and aging states. The methods and insights developed here support both academic research and industrial applications, offering practical guidelines for safer design and operation of large-format lithium-ion batteries in heavy-duty electric vehicles.
📄 Lab-scale versus industrial-scale thermal runaway tests for lithium-ion battery cells
Journal of Energy Storage, 2025,
Read more ↗
Ola Willstrand, Yang Yang et al.
Abstract: Lithium-ion battery safety is a topic of large importance, and testing is associated with large costs. Safety evaluation is therefore needed also at an early stage in cell design and development, in order to evaluate potential short-comings or to screen large platforms of materials and cells. Previous thermal runaway tests on lab-scale cells have, however, indicated differences in heat release and temperature increase as compared to commercial cells. On the other hand, these could also vary between different commercial cells due to differences in cell materials, size, format, and design. In this work, thermal runaway characteristics of industrial-scale cells are compared against each other as well as with lab-scale cells. Tests were performed on lab-scale coin cells and five different industrial-scale cells ranging from 5 to 157 Ah, covering different cell formats and materials. The coin cells were built using electrodes and separator extracted from one of the industrial-scale cells. The results show that the thermal runaway will be less violent and reach lower maximum temperatures using lab-scale cells, depending on the lower proportion of active materials as compared to inactive components. It is also shown that the ratio between cell capacity and heat capacity is a useful indicator for the comparability of thermal runaway scenarios. This ratio varies considerably for small cells, but not so much for different commercial cells. The data available suggests that a cell capacity of at least 1 Ah is needed to achieve a good comparison of the thermal runaway scenario with larger commercial cells.
📄 Investigating the effect of packing format on LiNixCoyMnzO2 lithium-ion battery failure behavior based on multidimensional signals
Journal of Power Sources, 2024,
Read more ↗
Yang Yang et al.
Abstract: Prismatic and pouch packaging formats are commonly used in LiNixCoyMnzO2 (NCM) batteries for electric vehicles, each showing distinct failure dynamics. However, a comprehensive study is lacking on how these packaging types affect thermal runaway (TR) at the cell level and its propagation at the module level, with a particular gap in understanding the dynamics of multidimensional signals. In this study, we experimentally explore the effect of cell format on 40 Ah NCM523 prismatic and pouch battery failure behaviors under overcharging and overheating conditions, by applying multidimensional signals, including the swelling force, gas, voltage, and temperature of the batteries. Results indicate that both types of batteries exhibit similar time scales for the failure modes when overcharged. In contrast, under overheating conditions, the pouch batteries fail significantly earlier than the prismatic batteries, including abnormal swelling, venting, gas emission, internal short circuit, and TR. Additionally, the prismatic batteries can withstand a swelling force of 5000 N at venting, while it is 2000 N for the pouch batteries. During TR, the prismatic batteries present a maximum temperature increase rate below 100 K/s, while the pouch batteries exhibit one over 200 K/s. Furthermore, the pouch batteries generally display more severe TR hazards and faster TR propagation than the prismatic cells. This study enhances the comprehension of TR and TR propagation mechanisms across different cell formats, providing crucial insights for the safety design and early warning strategies of battery modules.
📄 Investigating multidimensional signal evolution characteristics of LiFePO4 batteries under different thermal runaway scenarios
SSRN Preprint, 2024,
Read more ↗
Kuijie Li, Xinlei Gao, Yang Yang et al.
Abstract: Overheating and overcharging are two common trigger methods for thermal runaway (TR) in lithium-ion batteries, with distinct differences in failure behaviors in the multidimensional signal evolution. This study qualitatively and quantitatively investigates the multidimensional signal evolution, during TR and its propagation in battery samples at both the cell and the module levels, specifically focusing on the trigger methods of overcharging and overheating. The results demonstrate that the force anomaly can be identified at the earliest time sequence among these four signals (expansion force, gas concentration, temperature, and voltage), indicating that expansion force can effectively utilized as a warning signal for TR in a single cell, as well as for preventing TR propagation in the battery module under different conditions. At the cell level, under overcharging, the venting force and its rising rate are significantly higher, with force peaks reaching above 10000 N. On the other hand, overheating generates approximate 2 times higher venting temperature compared to overcharging. At the module level, adjacent batteries are found to be more susceptible to triggering propagation when subjected to overheating rather than overcharging. Especially, force characteristic peaks decease along with propagation direction. These findings highlight the influence of trigger methods on TR and failure propagation behaviors in multidimensional signal evolution. Consequently, they provide comprehensive insights for the development of effectively early warning strategy for battery TR.
📄 A cost-effective alternative to accelerating rate calorimetry: Analyzing thermal runaways of lithium-ion batteries through thermocouples
Journal of Power Sources, 2024,
Read more ↗
Yang Yang et al.
Abstract: In order to facilitate safety of heavy-duty batteries, approaches for studying thermal runaway (TR) need to be developed. So far, these have relied on accelerating rate calorimetry as a standard technique. This method, however, is costly, generally has size limitations, and is therefore of limited use for large format batteries. In this study, we examined the TR behavior of battery cells through a thermal propagation test at module level employing 157 Ah battery cells, using simple thermocouples. This constitutes one of the largest prismatic cell format analyzed to date, while the utilization of thermocouples enables a cost-effective method to study its TR. Parameters such as TR onset temperature, maximum temperature, heat release, and trigger time of the cells were comprehensively evaluated and compared, using this method. An onset temperature for TR at around 144 °C and a maximum temperature from 757 °C to 863 °C were observed. Heat release was estimated as 1.59 MJ per battery cell, deviating within ∼1 % compared to nail penetration tests. Moreover, six distinct stages during TR could be observed, in accordance with literature. This shows that the thermal propagation test using thermocouples is able to align well with other methods such as accelerating rate calorimetry, but is considerably easier to employ.
📕 Development of a Method to Measure Residual Stresses in Cast Components with Complex Geometries
KTH Library, 2019,
Read more ↗
Yang Yang
Abstract: Cast iron, taking the advantages of the advanced castability forming components of complex geometries and favorable mechanical properties, is employed in engine components in truck industries. Compacted graphite iron (CGI) integrates both merits of lamellar graphite iron (LGI) and spheroidal graphite iron (SGI) such as good machinability and high thermal conductivity from LGI, high ultimate tensile strength (UTS), good fatigue resistance, high elastic modulus, and high ductility from SGI, thus is now becoming a competitive alternative of traditional LGI in cylinder blocks and heads. Due to the shape complexity of cast components, residual stresses arise accordingly. Normal methods for measuring stresses have various practical difficulties that affect accuracy. For example, in strain gauge measurements such as hole drilling and cutting, casting skins need to be polished as the attachment of strain gauge requires a smooth surface condition for precise detection, though any mechanical treatment would change the residual stress state. On the other hand, electropolishing applied in XRD measurement for extracting depth profile causes no release of stresses, nevertheless, there is no dissolution reaction on graphite particles. This would retard further polishing and form a rough surface instead of flat extraction. A visual strain detection system relies on a stable and clean surface condition, therefore, when it is combined with the drilling technique, the drilling chips could be a vital problem for repeatability when they block the view of drilling edges. Ultrasonic measurement, in theory, has lower precision by averaging the stresses within a certain volume beneath surfaces. A number of methods have been developed to measure residual stresses, ranging from destructive to non-destructive according to the removal amount of materials. In this thesis work, several measurement methods are implemented on cylinder heads and the results are compared with simulation to develop a suitable method of measuring residual stresses in cast engine components. It is found that longer shakeout time lowers the tensile stresses and develops more compressive stresses in the surface layer. Cutting is a suitable method compared with others. Incremental center-hole drilling technique is not suitable to measure cast components as the surface grinding before stain gauge mounting causes high deviation. Hole drilling with visual strain detection provided high errors within the first 0.1 mm as the strains were too weak to be visualized at the beginning of drilling. The electropolishing process was also found retarded by graphite particles, and the XRD results are more trustworthy with more tilt angles. Ultrasonic measurement is rather rough due to the influence of graphite on the traveling velocity of ultrasound.
💡 An innovative solution for thermal propagation
Swedish Intellectual Property Office (PRV)，2025
LinkedIn
GitHub
Google Scholar
ORCID
ResearchGate
Email
Curriculum Vitae