A probabilistic occupant evacuation model for fire emergencies using Monte Carlo methods

The paper proposes a probabilistic occupant evacuation model for fire emergencies using Monte Carlo methods and then integrates the model into the fire risk analysis model CUrisk. The evacuation model uses a coarse network approach to describe a building and an individual perspective to represent occupants. The Monte Carlo methods repeat random variables including the occupants’ gender, age, speed, exit selection, pre-evacuation time, and distance to be traveled to exit the initial compartment with uniform distributions. The model produces results similar to those generated by deterministic models for cases dominated by deterministic factors, and gives results covering the possible ranges of outputs for cases controlled by random characteristics. The validation of the model against Test 9 in Guidelines for Evacuation Analysis for New and Existing Passenger Ships shows that predicted evacuation times are between the results given by Pathfinder 2009 and FDS+EVAC. They are also in agreement with the expectations for the test. The application of the model to the Station nightclub fire produces evacuation times covering or similar to those given by Pathfinder 2009, buildingEXODUS, and Simulex for the fire drill mode, and the percentages of remaining occupants in good agreement with the real fire consequences.

► A probabilistic occupant evacuation model for fire emergencies is proposed. ► The model uses Monte Carlo methods. ► The model is integrated into the fire risk analysis model CUrisk. ► The model produces good results when compared with other models or fire emergencies.