A reduced-order thermal runaway network model for predicting thermal propagation of lithium-ion batteries in large-scale power systems

Accurate and rapid prediction of thermal runaway propagation in a battery module and pack is essential for the thermal safety design and thermal runaway warning of large-scale lithium-ion battery power systems. This study introduces a highly accurate reduced-order thermal runaway network (TRN) model by redistributing heat source terms and correcting thermal runaway trigger criteria. Compared to traditional thermal network models, the TRN model attains precise simulation of thermal runaway propagation, ensuring the accuracy of thermal runaway trigger time to within 1 min. Subsequently, the effectiveness of the model is demonstrated by simulating the thermal runaway propagation process in a commercial battery pack. The findings indicate an accelerating trend in thermal propagation due to the heat accumulation effect once thermal runaway initiates within a module. Specifically, the thermal runaway propagation interval is markedly reduced by 72%, from 611 s to 176 s, indicating that controlling the initial thermal spread within the module is more critical than mitigating thermal propagation between modules. Moreover, during thermal runaway propagation, the energy dissipated via the liquid cooling plates accounts for more than 70% of the total energy, far exceeding the heat transfer through neighboring battery surfaces and the tab connector. The TRN model with high accuracy and reliability can facilitate the development of onboard thermal runaway warning systems and provide valuable insights for thermal runaway inhibition, guiding the thermal safety design of lithium-ion battery power systems.