A Review of Building Evacuation Models - NIST Virtual Library

parameters yields realistic results, as already demonstrated by the investigation of other
evacuation drills.” For the tallest building, a 21-story office building with 1400 occupants, the
calculated total evacuation times ranged between 616 s and 648 s, with a mean value of 627 s,
while the measured evacuation time for the structure was 629 s. More information on this
validation case study is provided in ASERI references.
The final case study to be discussed in this section involved the evacuation from a hotel
conducted by the Norwegian SINTEF organization. The input information provided to the
model for this case study involved the building layout, means of egress, geometrical staircase
information, location and the sequence of the fire incident, and the communication events put in
place by the evacuation plan. The evacuation case that follows the evacuation plan is called the
“schedule case” and actual observation of the drill is referred to as the “actual case.” Also,
information about the occupants was available such as the gender, age, room number, and
activity engaged in before evacuation began. The staff was not included in the egress movement
during the simulation, but was modeled to perform actions during the alarming sequence. Also,
delay and response times associated with certain occupant actions were included in the
simulation. The occupant total was 104, and since the available egress routes were many, the
evacuation was not influenced by crowding. As mentioned earlier, runs were performed in
ASERI to simulate 1) immediate evacuation of all occupants at the start of the fire alarm, 2) the
scheduled case, and 3) the actual case. According to the developers, the actual case was very
much in agreement with the observation of the monitored hotel drill. The only difference noted
was that “the number of occupants not leaving the guest rooms or returning into the room was
much larger than predicted by the simulation.” The developers relate this discrepancy to the fact
that the information available was ambiguous in the drill, resulting in guests ignoring the alarm.
Additional validation studies can be found in the referenced ASERI publications.
Special features:
Manual exit block/obstacles – Yes, if smoke is very heavy.
Fire conditions affect behavior Yes, the output of KOBRA-3D can be transferred to ASERI
through a cut and paste method.
Defining groups – Yes, because of the ability to assign each individual certain mobility
parameters (body size, walking speed, and behavioral conditions) as well as providing a
distribution of these for a specified group.
Disabilities/slow occupant groups – Yes, walking speed and increased body size can be
specified.
Delays/pre-movement time – Delays are achieved either by assigning alarm and reaction times or
introducing intermediate stop positions.
Toxicity of the occupants – Yes.
A-38