Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Testing the crowd:it model for building fire evacuation scenarios

Movitz Rydhé, Niclas LU (2024) In LUTVDG/TVBB VBRM01 20222
Division of Fire Safety Engineering
Abstract
Evacuation models can be reliably used if they are sufficiently validated for the application in
question (e.g. Fire Safety Engineering). To be useful, an evacuation model should therefore be an
accurate enough representation of the real-world behaviour of evacuees. The scope of this work
was to determine how accurately the model crowd:it (making use of the Optimal Steps Model)
provides a representation of the real-world behaviour in building fire evacuation. Validation
tests were conducted in accordance with the methodology of verification and validation protocol
for building fire evacuation models presented in ISO 20414. Suggested tests for assessing parts
of the model representing pre-evacuation time, effect on walking speed at... (More)
Evacuation models can be reliably used if they are sufficiently validated for the application in
question (e.g. Fire Safety Engineering). To be useful, an evacuation model should therefore be an
accurate enough representation of the real-world behaviour of evacuees. The scope of this work
was to determine how accurately the model crowd:it (making use of the Optimal Steps Model)
provides a representation of the real-world behaviour in building fire evacuation. Validation
tests were conducted in accordance with the methodology of verification and validation protocol
for building fire evacuation models presented in ISO 20414. Suggested tests for assessing parts
of the model representing pre-evacuation time, effect on walking speed at different densities,
movement on stairwells and in flight of stairs, counter-flow, route/exit choice, and movement in
bottlenecks at openings were conducted. Two global tests of full evacuations of an auditorium
and a school were performed as well. There were a few tests where behaviours differed to a
greater extent. In counterflow conditions, agents tended to get stuck, and agents tended to zigzag between lanes in a flow rather than following the person in front (e.g. lane formation
phenomena) as experiments showed. Ascending or descending a stairwell does not alter the
movement speed of agents when stairs are modelled explicitly. They can however also be
implicitly modelled similar other evacuation models. Distributions available for representing
evacuees’ characteristics and behaviour in crowd:it was recommended to be expanded and the
log-normal distribution was added due to findings in this report. In summary, validation test
results were generally in line with the experimental results and verification tests highlighted a
set of model features that could be improved. (Less)
Please use this url to cite or link to this publication:
author
Movitz Rydhé, Niclas LU
supervisor
organization
course
VBRM01 20222
year
type
M2 - Bachelor Degree
subject
keywords
Evacuation modelling, Validation, ISO 20414, crowd:it, Optimal Steps Model
publication/series
LUTVDG/TVBB
report number
5716
other publication id
LUTVDG/TVBB--5716--SE
language
English
id
9151160
date added to LUP
2024-04-30 08:35:33
date last changed
2024-04-30 08:35:33
@misc{9151160,
  abstract     = {{Evacuation models can be reliably used if they are sufficiently validated for the application in
question (e.g. Fire Safety Engineering). To be useful, an evacuation model should therefore be an
accurate enough representation of the real-world behaviour of evacuees. The scope of this work
was to determine how accurately the model crowd:it (making use of the Optimal Steps Model)
provides a representation of the real-world behaviour in building fire evacuation. Validation
tests were conducted in accordance with the methodology of verification and validation protocol
for building fire evacuation models presented in ISO 20414. Suggested tests for assessing parts
of the model representing pre-evacuation time, effect on walking speed at different densities,
movement on stairwells and in flight of stairs, counter-flow, route/exit choice, and movement in
bottlenecks at openings were conducted. Two global tests of full evacuations of an auditorium
and a school were performed as well. There were a few tests where behaviours differed to a
greater extent. In counterflow conditions, agents tended to get stuck, and agents tended to zigzag between lanes in a flow rather than following the person in front (e.g. lane formation
phenomena) as experiments showed. Ascending or descending a stairwell does not alter the
movement speed of agents when stairs are modelled explicitly. They can however also be
implicitly modelled similar other evacuation models. Distributions available for representing
evacuees’ characteristics and behaviour in crowd:it was recommended to be expanded and the
log-normal distribution was added due to findings in this report. In summary, validation test
results were generally in line with the experimental results and verification tests highlighted a
set of model features that could be improved.}},
  author       = {{Movitz Rydhé, Niclas}},
  language     = {{eng}},
  note         = {{Student Paper}},
  series       = {{LUTVDG/TVBB}},
  title        = {{Testing the crowd:it model for building fire evacuation scenarios}},
  year         = {{2024}},
}