Elsevier

Safety Science

Volume 47, Issue 1, January 2009, Pages 97-104
Safety Science

The incorporation and validation of empirical crawling data into the buildingEXODUS model

https://doi.org/10.1016/j.ssci.2007.12.003Get rights and content

Abstract

The deterioration of environmental conditions can influence evacuee decisions and their subsequent behaviors. Simulating evacuee behaviors enhances the robustness of engineering procedural designs, improves the accuracy of egress models, and better evaluates the safety of evacuees. The purpose of this paper is to more accurately incorporate and validate evacuee crawling behavior into the buildingEXODUS egress model. Crawling data were incorporated into the model and tested for accurate representation. Once confirmed, the data were compared to another crawling data set for validation. The buildingEXODUS model demonstrated the ability to incorporate and compare the effects of crawling behavior. The paper further suggests some enhancements to be made to models that attempt to simulate crawling.

Introduction

Safety engineers are faced with increasingly sophisticated egress models to evaluate the safety of evacuees in a wide range of building designs and structures. Invariably, egress models require a comprehensive representation of human behavior when faced with emergency egress, to better establish the effectiveness of these behaviors and their impact on the overall evacuation process. This recently developed approach tests our understanding of egress behavior. Therefore, there is a need to develop new methods or enhance existing ones in order to cope with the level of detail required to ensure evacuee safety.

Ideally, the successful application of emergency egress procedures should ensure the protection of the evacuating population from the deteriorating environmental conditions. However, such protection may not be guaranteed in some engineering applications and incident scenarios. For instance, a forensic analysis could be conducted where the impact of certain types of behavior need to be established under severe fire conditions. Further, a design may be scrutinized according to a number of fire scenarios (varying in severity) where the evacuating population is forced to interact with fire. As a result, it is important to understand the robustness of designs and the consequences of the population encountering such environment. Establishing the consequences of implementing procedural designs given specified fire scenarios, as represented by the application of egress models, is therefore of great value to safety engineers.

The physiological effects of exposure to fire outcomes influence the ability of evacuees to make decisions and perform behavioral actions. Bryan (1977) and Wood (1980) reported that evacuees move through, or redirect away from smoke. Jin (1976) suggested that evacuees may stagger through smoke and subsequently be forced to move along the perimeter of the building to guide their movement. Evacuees may also respond to a descending hot layer of smoke by crawling.

The ability to simulate these behaviors, and others, will enhance the robustness of engineering designs, improve the accuracy of egress models, and ultimately better assess evacuee safety. The purpose of this paper is to demonstrate a method for increasing the accuracy of an existing egress model, namely buildingEXODUS (Gwynne et al., 1999, Jiang et al., 2003, Park et al., 2003, Galea et al., 2004) by simulating evacuee crawling using empirical data; i.e. by improving the existing crawling functionality through the use of empirical data. By doing so, the model can be applied to examine the outcome of evacuee crawling allowing more informed procedural designs and forensic analyses. Therefore, it is important to accurately represent the speeds that can be maintained while crawling to assess the consequences of such behavior. Crawling speeds may directly influence the survival rate, and also have an impact on the surrounding population by leading to congestion.

Section snippets

The approach

In order to enable and validate the accurate representation of crawling speeds within the buildingEXODUS computational egress model, two separate crawling data sets are employed. The Muhdi et al. (2006) data are utilized within the model to determine the achievable individual crawling speeds during the simulation. The end results, in terms of overall egress times, are then compared against the Nagai et al. (2006) data. The representation of crawling is then improved allowing greater confidence

Design assessment techniques in fire scenarios

Two methods are usually employed to determine whether a design affords sufficient protection to the resident population: applying prescriptive codes and conducting a performance-based evaluation. Applying prescriptive codes requires pre-defined guidelines to be followed when designing structural components. The overall design is then deemed to reach a sufficient level of safety without the need for further analysis. For instance, it is not possible or necessary to establish the level of safety

The incorporation of fire data into egress models

Incorporating fire data directly into egress models (by coupling the outcome of the fire conditions and the simulated evacuation) requires fewer scenario assumptions (e.g. the population is never exposed to the fire conditions). However, the decision to include such data in the models poses several challenges to safety engineers

  • How are the data represented in the model?

  • How does the model represent the physical characteristics of an evacuee?

  • How does the presence of fire conditions affect evacuee

Description of crawling data

Muhdi et al. (2006) measured and compared individual maximum walking and normal and maximum crawling to normal walking speed. Twenty six students participated in the study. The subjects were familiar with the study prior to conducting the protocal. The physical activities required in the study were thoroughly explained and practiced. The test area was 100 ft long, marked every 20 ft. The start and finish lines were set about 10 ft from the beginning and the end of the test area, respectively,

Description of the egress model buildingEXODUS

The buildingEXODUS model is a suite of software tools designed to simulate the evacuation of large numbers of people from a variety of enclosures. It is comprised of five core interacting sub-models: the Occupant, Movement, Behavior, Toxicity and Hazard. The software describing these sub-models is rule-based, with the progressive motion and behavior of each individual being determined by a set of heuristics. On the basis of an individual’s personal attributes, the Behavior sub-model determines

Inclusion of the Muhdi data into the buildingEXODUS model

The first step in the analysis was to incorporate the Muhdi data into buildingEXODUS by employing the reduction factors calculated for both normal and maximum speeds. This represents the effective reduction in travel speeds caused by crawling. The speed reduction approach gives a more reliable representation to the relationship between crawling and walking speeds than absolute travel speeds from the empirical data, especially in situations where evacuees with a variety of capabilities travel at

Results

The accurate incorporation of the Muhdi data was first confirmed through the use of several simple tests, prior to comparing any results with the Nagai data. In these simple cases, an individual was forced to crawl for 15 m to reach an exit. The results produced by buildingEXODUS, shown in Table 4, confirm the accurate representation of the data within the model. Once the successful incorporation of the data had been confirmed, the comparisons could then be made between the simulated and

Discussion

In this paper, an attempt has been made to examine the impact of applying new empirical crawling speed data within an evacuation simulation model. The buildingEXODUS model was selected as the test platform, given its ability to represent crawling. The default buildingEXODUS crawling reduction factor and the Muhdi data were incorporated into the model. The results were then compared against the Nagai group experimental data set. The employment of the Muhdi data consistently produced results

Conclusions

This paper has demonstrated the impact of incorporating empirical crawling data into buildingEXODUS; it has also demonstrated the validity of this approach through comparison with a second empirical data set. The Muhdi crawling data were incorporated and the results produced were compared against the Nagai’s group experiments; the results produced were favorable. The inclusion of this data has therefore been shown to improve the existing functionality of the model, and to provide a credible

References (26)

  • D.A. Purser

    Behavioural impairment in smoke environments

    Toxicology

    (1996)
  • G.N. Berlin

    A simulation model for assessing building fire safety

    Fire Technology

    (1982)
  • J.L. Bryan

    Smoke as a Determinant of Human Behavior in fire Situation

    (1977)
  • J. Ewer et al.

    SMARTFIRE V3.0 User Guide and Technical Manual

    (2002)
  • R. Fahy

    Exit 89 – An evacuation model for high-rise buildings: Model description and example applications

  • J.N. Fraser-Mitchell

    An object oriented simulation (CRISP II) for fire risk assessment

  • Galea, E.R., Gwynne, S., Lawrence, P.J., Filippidis L., Blackshields, D., Cooney, D., 2003. buildingEXODUS version 4.0:...
  • Galea, E.R., Gwynne, S., Lawrence, P., Filippidis, L., BlackShields, D., Cooney, D., 2004. buildingEXODUS Technical...
  • S. Gwynne et al.

    Adaptive decision making in response to crowd formations in buildingEXODUS

    Journal of Applied Fire Science

    (1999)
  • Heskestad, A.W., Meland, Ø.J., 1998. Determination of evacuation times as a function of occupant and building...
  • H. Jiang et al.

    The use of evacuation simulation, fire simulation and experimental data in forensic fire analysis

  • Jin, T., 1976. Visibility through fire smoke. Report of Fire Institute of Japan #42, Tokyo,...
  • Kuligowski, E.D., Peacock, R.D., 2005. A review of building evacuation models. NIST Technical Note 1471, U.S. Dept. of...
  • Cited by (7)

    • A review of optimisation models for pedestrian evacuation and design problems

      2016, Safety Science
      Citation Excerpt :

      EXODUS is a similar software tool, developed by Galea and Perez Galparsoro, intended for the evacuation of mass-transport vehicles such as aircraft (Galea and Perez Galparsoro, 1994). It is also able to simulate crawling movement during evacuations (Muhdi et al., 2009). A third software tool, developed for simulating the evacuation of geometrically complex buildings, is the SIMULEX model of Thompson and Marchant (1995a,b,c).

    • The development of a movement-density relationship for people going on four in evacuation

      2012, Safety Science
      Citation Excerpt :

      A crawling speed of 0.3 m/s (0.98 ft/s) is significantly less than the findings of Muhdi et al. (2006) and Nagai et al. (2006); i.e., 0.71 (2.33 ft/s) and 0.73 m/s (2.4 ft/s), respectively. In an attempt to incorporate crawling data into evacuation models, Muhdi et al. (2008) employed the crawling data found in Muhdi et al. (2006) and Nagai et al. (2006) into buildingEXODUS (Galea and Galparsoro, 1994; Owen et al., 1996) to test the accurate representation of crawling speeds in the model. The researchers emphasized the importance of incorporating reliable occupant data into evacuation models to verify the model outcomes, and suggested the development of a density–speed relationship for crawlers to improve upon the model’s representation of crawling movement.

    • Effect of height constraints on unidirectional pedestrian flow

      2022, Journal of Statistical Mechanics: Theory and Experiment
    • A VR-based, hybrid modeling approach to fire evacuation simulation

      2018, Proceedings - VRCAI 2018: 16th ACM SIGGRAPH International Conference on Virtual-Reality Continuum and its Applications in Industry
    View all citing articles on Scopus
    View full text