Application of influence diagrams to multi‐objective allocation of firefighting resources in process plants

Abstract Identification of firefighting strategies (i.e., which endangered units to suppress or cool first) in chemical and process plants falls under the domain of multi‐objective decision‐making (MODM), where not only the safety and integrity of the affected process plant but also the safety of on‐site and off‐site vulnerable targets matter. The importance of identifying effective firefighting strategies becomes more crucial when potential domino effects (i.e., escalation of a primary fire to secondary fires) can quickly increase the number of endangered units and targets beyond what could initially be handled by firefighters. While modelling and risk assessment of domino effects have gained increasing attention over the past decade, developing methods for risk management and firefighting of domino effects has lagged. In this regard, combining the domino effect models with MODM techniques has been proposed as a viable solution for identifying effective firefighting strategies. In the present, by developing an innovative multi‐attribute utility function (MAUF), it will be shown that influence diagrams—an extension of Bayesian networks—can be applied both for modelling domino effects and for identifying multi‐objective firefighting strategies within an integrated framework. The results of the developed method are shown to be consistent with those obtained from other MODM techniques, such as goal programming.