A graph-based optimization methodology for identifying firefighting strategies aimed at domino effects

Firefighting strategies at process plants would include simultaneous extinguishment of burning units and cooling of exposed units if firefighting resources are sufficient. This way, the fire can be contained and its propagation to the exposed units can be prevented, which would otherwise cause fire escalation and result in domino effects. However, when the firefighting resources are not sufficient to handle all the critical units—either burning or exposed—at once, firefighters need to decide which burning units to suppress first and which exposed units to cool first to minimize the risks. Making effective decisions in such situations becomes critical knowing that, by spreading the fire to adjacent units, the number of critical units grows exponentially, making the available firefighting resources even more insufficient. In the present study, after modelling fire spread in a tank terminal as a directed graph, closeness centrality—a graph centrality metric—is used to identify the critical units from the viewpoint of their contribution to potential domino effects. Knowing the critical units and considering available firefighting resources for suppression and cooling of these units, mathematical programming is applied for optimal allocation of firefighting resources. A comparison between the results of the present work and previous studies shows the effectiveness of the developed methodology.