Abstract
In the event of major fires in process plants, effective emergency response becomes highly challenging for firefighters. Ideally, the burning tanks should be suppressed, and the exposed tanks should be cooled to prevent fires from escalating and the onset of domino effects. When available resources are limited, prioritization becomes essential for firefighters as the domino effect can increase the number of threatened units beyond the operational capacity of firefighters. To address this scenario, the present study introduces a decision-making method based on Bayesian network and game theory to capture complexities and uncertainties arising from the sequential nature of firefighting actions and the dynamics of the domino effect. To do so, the interactions between firefighters and fire have been modeled via an extensive game, while Bayesian network has been used to model domino effects and calculate the payoffs needed in the game modeling. This approach is demonstrated to provide a more realistic representation of how priorities should be established during major fires in process plants and how firefighters should react to changing fire scenarios.
