Modeling the variability of thermal runaway behavior in large-format cylindrical cells with NMC-811 and LFP cathodes: A simulation study
Abstract
Accurate thermal runaway modeling is required to mitigate thermal runaway propagation in battery systems. Since the thermal runaway of battery cells is a chaotic process, it is subject to high statistical variance. However, the variability of thermal runaway behavior is often not reflected in current simulation models. This paper presents a methodology for modeling the variability of thermal stability and heat release during thermal runaway of NMC-811 and LFP battery cells using accelerating rate calorimetry (ARC) and discretized autoclave thermal runaway calorimetry (DATRC) experiments. The thermal runaway model validation demonstrates a good agreement between simulation and experimental data within one standard deviation. In a simulation study, the thermal runaway simulation model was extended to a thermal runaway propagation model for a battery system with cylindrical 4680 cells. The simulation study revealed that thermal runaway propagation by heat transfer is unlikely in the case of a LFP cell chemistry. In contrast, thermal runaway propagation in similar systems with NMC-811 cells strongly depends on mass loss, cell spacing, and housing material. The simulation model improves an accelerated safety design concerning thermal runaway propagation in battery systems and enables a potential error estimation. Future research should focus on transferring the approach to other cell formats and validation on the system level. • Thermal runaway simulation framework for cylindrical NMC-811 and LFP cells. • Modeling the variability of thermal runaway behavior in lithium-ion batteries. • Model-based safety-related design of battery systems with cylindrical battery cells.
Fields of Study & Subjects
Advanced Battery Technologies Research
Advanced battery technologies research
Reliability and Maintenance Optimization
Publication details
- Journal / source
- Journal of Power Sources
- Publisher
- Elsevier BV
- Volume
- 668
- Pages
- 239206–239206
- Publication date
- 22 January 2026
- Language
- EN