Fire Risk Assessment of An Energy Storage Station Based on Bayesian Network and PyroSim

Lithium-ion battery storage stations have become a crucial component of modern power systems, yet their inherent instability poses severe fire risks during storage. Existing research primarily addresses post-fire control and suppression, leaving a gap in understanding the comprehensive factors influencing fire risk before and after an incident. This study aims to bridge this gap by proposing an integrated risk assessment method combining Bayesian networks and PyroSim. The Bayesian network is employed for both qualitative and quantitative analysis of key factors contributing to fire risk in storage stations, identifying high-risk nodes. PyroSim software is then used to simulate fire scenarios based on these critical risk factors, enabling dynamic analysis of smoke dispersion and heat transfer throughout the fire process. Our findings reveal a substantial $\mathbf{6 9 \%}$ probability of fire originating from thermal runaway in battery modules. Moreover, with an operational fire suppression system, internal fire temperatures can be reduced by up to $400^{\circ} \mathrm{C}$. This integrated approach enhances understanding and management of fire risks in lithium-ion battery storage stations, providing valuable insights for improving safety protocols.