Abstract: In order to solve the problem that the Bow-tie model cannot quantify the risk of LNG storage tanks, a Bow-tie model was established with five high-risk elements of LNG storage tanks: high-risk sites, high-risk processes, high-risk equipment, high-risk substances, and high-risk personnel. Then the normal cloud model is introduced to establish a cloud model for the cloud drop based on the possibility and severity of risk. Combining with Python language visual program, a cloud map was generated and the cloud closeness was calculated to determine the risk level of each element. Finally, analysis was carried out with the example of an enterprise LNG storage tank farm. The results show that the high-risk sites in the Bow-tie model are the first-level element with the highest risk, and the LNG storage tank pre-cooling process is the second-level element with the highest risk. The two basic events of pressure and temperature are the keys to LNG leakage. The installation of pressure-temperature automatic sensing system can effectively prevent and control accidents. It provides a basis for dynamic risk management and control of storage tanks.

Key words: LNG storage tank, Bow-tie model, normal cloud model, Python visualization, risk analysis