Vulnerability of industrial plants to flood-induced natechs: A Bayesian network approach

In the context of natural-technological (natech) accidents, flood-induced damage of industrial plants have received relatively less attention mainly due to the scarcity of such accidents compared to those triggered by earthquakes, high winds, and lightnings. The large amount of oil and chemicals spillage due to floods triggered by the Hurricanes Katrina and Rita in 2005 and Harvey in 2017 in the U.S. demonstrated the potential of floods in causing catastrophic natechs. In the present study, we have developed a methodology based on physical reliability models and Bayesian network so as to assess the fragility (probability of failure) of industrial plants to floods. The application of the methodology has been demonstrated for petroleum storage tanks where flotation, shell buckling, and sliding are considered as the prevailing failure modes. Due to scarcity of empirical data and high-resolution field observations prevailing in natechs, the developed methodology can effectively be applied to a wide variety of natechs in industrial plants as long as limit state equations of respective failure modes can reasonably be developed.