Abstract: To support the comprehensive risk assessment of chemical industrial areas and address the difficulties of quantifying uncertainties related to complex spatiotemporal evolution of coupled multi-hazard, a systematic study of risk factors, evolution laws, and dynamic probability analysis method of coupled multi-hazard in the chemical tank farm is carried out. First, the disaster system in the chemical tank farm and related risk factors are explored from the perspective of equipment and tank farm. Second, the evolution pattern of coupled multiple hazards in the chemical tank farm is depicted using multi-level Bow-Tie structures, providing theoretical basis for the establishment of accident evolution network. Finally, the dynamic Bayesian network is used as a tool to quantify dynamic uncertainties, and a dynamic probability analysis method for coupled multi-hazard in the chemical tank farm is developed. The results of case study show that natural hazard factors could lead to the rise of tank failure probability. Moreover, the key unit for accident evolution can be identified from the curves of dynamic probability. The developed method could provide data support for risk analysis and emergency management of coupled disasters in any chemical industrial area.

Key words: chemical tank farm, coupled multi-hazard, spatiotemporal evolution, dynamic Bayesian network, dynamic probability