A review of the methodologies used in the computer simulation of evacuation from the built environment
S. Gwynne,
Fire Safety Engineering
Received
E.R. Galea *, M. Owen, P.J. Lawrence,
Group, Centre,for
26 February
Numerical
Mode&g
and Process Analysis,
1998; received in revised form 14 September
L. Filippidis
Uniuersit~~ qf Greenwich,
1998; accepted
7 October
London, U.K.
1998
Abstract
Computer
based analysis of evacuation can be performed using one of three different approaches, namely optimisation, simulation or risk assessment. Furthermore, within each approach different means of representing the enclosure, the population, and the behaviour of the population are possible. The myriad of approaches which are available has led to the development of some 22 different evacuation models. This article attempts to describe each of the modelling approaches adopted and critically review the inherent capabilities of each approach. The review is based on available published literature. 0 1999 Elsevier Science Ltd. All rights reserved. 1. Introduction
As architects continue to implement novel concepts in building design, they are increasingly faced with the dilemma of demonstrating in some manner that their concepts are safe and that the occupants will be able to efficiently evacuate in the event of an emergency. Traditionally, two techniques have been used to meet these needs: (1) full-scale evacuation demonstration, and (2) the adherence to prescriptive building codes.
The full-scale evacuation demonstration involves staging an evacuation exercise using a representative target population within the structure. Such an approach poses considerable ethical, practical and financial problems that bring into question its viability.
The ethical problems concern the threat of injury to the participants and the lack of realism inherent in any demonstration evacuation scenario. As volunteers cannot be subjected to trauma or panic nor to the physical ramifications of a real emergency situation such as smoke, tire and debris, such an exercise provides little useful information regarding the suitability of the design in the event of a real emergency.
On a practical level, when evacuation drills are performed, usually only a single evacuation trial is undertaken.
Thus there can be limited confidence that the testwhether successful or not-truly represents the evacu-
*Corresponding author.
Tel.: +44
8925
0360-l 323/99/$ - see front matter
PII: SO360-1323(98)00057-2
0
181 331 8730; fax: +44
181 331
ation capability of the structure.
In addition, from a design point of view, a single test does not provide sufficient information to arrange the layout of the structure for optimal evacuation efficiency.
The need to perform repeated experiments should come as no surprise as even under the most controlled experimental conditions, no evacuation exercise involving crowds of real people will produce identical results if the exercise is repeated-even if the same people are used.
Hence it is unwise to make definitive statements such as
‘the evacuation time for the structure will be 187.7 s’ on the basis of a simple one off experimental analysis. For any structure/population/environment combination, the evacuation performance of the combination is likely to follow some form of distribution, a purely hypothetical example of such a distribution is provided in Fig. 1 (readers should draw no inference from the actual shape of the depicted distribution).
A single observation of evacuation performance could fall anywhere on the curve.
However, what can be achieved is an understanding of how the structure/population/environment system is likely to behave given a set of pre-defined conditions. Hence, for a given building configuration, specified type of occupancy and specific type of scenario, it is necessary to determine the range of evacuation performance likely to be achieved.
Finally, to
