Simulation modeling of a statistical fire spread to respond fire accident in buildings

Abstract

A statistical modeling of fire spread is intended to enable to firelighters to make the best plan by expecting the fire spread and to effectively extinguish a fire in buildings. In this Fpaper, Anylogic simulation modeling of a statistical fire spread is introduced for fire accident correspondence in buildings. For the modeling, we apply statistical modeling methods, such as Bayesian network and decision tree, to predict the fire spread. In addition, several conditions, including the fire growth parameter, barrier failure probability, the fire ignition location and an exceptional condition such as backdraft, are apply to the modeling methods. The modeling result allows firefighters to realize where the fire is located and how the fire is growing. Especially if the modeling can be applied to easily viewable modeling software such as AnyLogic, the firefighters can recognize the fire spread and growth in a map quickly, exactly and economically. Thanks to the modeling, firefighters can extinguish the fire easier and mortality rate and property damages, as well as the flrefighting time, can be decreased.