VULNERABILITY ANALYSIS OF THE TANKS OF AN OIL REFINERY TO FIRE-INDUCED DOMINO EFFECTS BASED ON GRAPH THEORY

Tank farms are vulnerable to fire-induced domino effects due to the high amount of flammable materials. The present study tries to analyze the vulnerability of the crude oil tanks in one of the refineries of Iran to domino fires. The purpose of vulnerability analysis is to determine the tanks with the greatest potential to initiate and propagate domino effects in order to manage the risk of domino fires.

Methods and Materials: 

Graph theory and centrality measures (betweenness and closeness) were used to model domino scenarios and analyze the vulnerability of tanks to domino effects. ALOHA consequence modeling software was used to model the consequences of the accident scenarios. R-statistical software and igraph package were used to model the graph and calculate centrality measures.

The possible domino scenarios in the tanks were modeled as a directed graph and the tanks were ranked based on the value of betweenness, out-closeness and in-closeness of their respective nodes (which represent the potential for the propagation of domino effects, potential for initiating domino effects and potential to be affected by other tanks during domino effects respectively). Also, based on all-closeness measure, the most probable sequence of tank involvement in a domino fire was determined.

Using graph theory allows the graphical modeling of domino scenarios and determining the most vulnerable tanks in domino scenarios. The results of vulnerability analysis using graph metrics can be used in the field of domino effects risk management and the most vulnerable tanks can be prioritized to assign protection measures