Validation of premovement time distribution for evacuation modelling using real data from evacuation drills

The primary component of a performance-based fire safety analysis is the prediction of evacuee movement. The forecast of how people respond to building emergencies can aid in safer building design, improving the level of safety afforded by these designs and reflecting its benefits also in the development of more effective emergency procedures, emergency communication systems, and pre-event emergency training for buildings and communities. Therefore, the use of evacuation models imply the difficult choice of parameters that govern human behaviours, since these parameters implicitly have a large uncertainties. This paper aims to compare the results of an evacuation drill involving a large number of people with agent-based evacuation models, using FDS+Evac and Pathfinder. Nevertheless, a plausible estimation of the pre-movement activity time is a challenge in the field of fire safety of buildings and the possibility of comparing evacuation drills experimental data to the results of evacuation models gives importance to the topic.