Performance-Based Design of a Hotel Building Using Two Egress Models: A Comparison of the Results.
Performance-Based Design of a Hotel Building Using Two
Egress Models: A Comparison of the Results.
(292 K)
Kuligowski, E. D.; Milke, J. A.
Human Behavior in Fire: Public Fire Safety -
Professionals in Partnership. International Symposium,
3rd. Proceedings. September 1-3, 2004, Belfast, N.
Ireland, Interscience Communications Ltd., London,
England, 399-410 pp, 2004.
Keywords:
human behavior; fire safety; decision making; hotels;
egress; evacuation; computer models; evacuation time;
fire protection engienering; scenarios; human response
Abstract:
With the move toward performance-based design, engineers
have been looking to evacuation computer models to
assess a building's life safety. Many times, the
engineer is tasked with the selection of one evacuation
model for a specific project. Currently there is a wide
variety of evacuation models for engineers to choose
from. However, with each model containing its own unique
features and simulation capabilities, confusions may
arise as to which model is best for the task at hand.
The results gained from this work emphasize the
importance for users to choose an egress model for each
project with the appropriate input features and
simulation capabilities. Current evacuation models have
the capability of simulating many of the complex
behaviors associated with a hotel building evacuation,
however, frequently without sufficient behavioral data
as a basis. Because of this, model users may choose
models containing less behavioral sophistication for
their performance-based design. However, model choice is
an important step, even among models with less
behavioral sophistication. This report compares results
from two similar egress models based on documented
evacuation movement data. When EXIT89 and Simulex (both
only partial-behavioral models) are used to 1) simulate
the same design scenarios and 2) perform a bounding
analysis of the hotel building, significant differences
in egress times were identified. EXIT89's evacuation
times were found to be 25-40% lower than Simulex for the
design scenarios, attributed to differences in unimpeded
speeds, movement algorithms, methods of simulating slow
occupants, density in the stairs, and stair
configuration input between the models. For the bounding
analysis, EXIT89 produced maximum evacuation times
30-40% lower than Simulex, primarily due to the
simulation of slower moving occupants.
Building and Fire Research Laboratory
National Institute of Standards and Technology
Gaithersburg, MD 20899