Hurricane as one of the most destructive natural hazards can make a devastating impact on the industrial equipment, especially atmospheric storage tanks, leading to the release of stored chemicals and disastrous safety and environmental issues. These catastrophic consequences are caused not only by strong winds but also by the torrential rainfall and inundating floods. The objective of this study is to present a risk-based methodology for assessing and reducing the vulnerability of atmospheric storage tanks to hurricanes. Considering the shell buckling, flotation, sliding, and roof sinking as dominant failure modes of atmospheric storage tanks during hurricanes, Bayesian network (BN) has been employed to combine the failure modes while considering their conditional dependencies. The probability updating feature of the developed BN was employed to indicate that the flood is the most critical hazard during hurricanes while the impact of wind and rainfall cannot be neglected. Extending the developed BN to an influence diagram, the cost-benefit filling of storage tanks with water prior to the advent of hurricanes was shown as a viable measure for reducing the damage probability. The results show that the proposed methodology can be used as an effective decision support tool for assessing and reducing the vulnerability of atmospheric storage tanks to natural hazards.

飓风是最具破坏性的自然灾害之一，会对工业设备（尤其是大气储罐）产生毁灭性影响，导致释放储存的化学物质以及灾难性的安全和环境问题。这些灾难性后果不仅是由强风引起的，而且还由暴雨和洪水泛滥引起。这项研究的目的是提出一种基于风险的方法来评估和减少大气储罐对飓风的脆弱性。考虑到壳体屈曲，漂浮，滑动和顶板下沉是飓风期间大气储罐的主要失效模式，在考虑其条件依赖性的情况下，采用贝叶斯网络（BN）组合失效模式。所开发的BN的概率更新功能用于表明洪水是飓风期间最严重的危害，而风和降雨的影响却不能忽略。将已开发的BN扩展到影响图，显示出在飓风来临之前向水箱中注水的成本效益是降低损坏可能性的可行措施。结果表明，所提出的方法可以用作评估和减少大气储罐对自然灾害的脆弱性的有效决策支持工具。事实证明，在飓风来临之前用水填充储罐的成本效益是降低损坏可能性的可行措施。结果表明，所提出的方法可以用作评估和减少大气储罐对自然灾害的脆弱性的有效决策支持工具。事实证明，在飓风来临之前用水填充储罐的成本效益是降低损坏可能性的可行措施。结果表明，所提出的方法可以用作评估和减少大气储罐对自然灾害的脆弱性的有效决策支持工具。