Open Access System for Information Sharing

Login Library


Cited 0 time in webofscience Cited 0 time in scopus
Metadata Downloads

High-rise Building Evacuation Problem with Heterogeneous Population including the Disabled

High-rise Building Evacuation Problem with Heterogeneous Population including the Disabled
Date Issued
This dissertation addresses a simulation study on the evacuation of a heterogeneous population, including people with disabilities and those without disabilities, from a 24-story high-rise building. The efficient evacuation of high-rise buildings is important to prevent human casualties during catastrophic events. Therefore, many researchers have attempted to make their evacuation models realistic, to reflect more of the phenomena that could occur in real emergencies. However, most evacuation studies do not consider a heterogeneous population, even though people with disabilities comprise between 10 percent and 20 percent of the population in many countries. In this dissertation, we study the evacuation of a heterogeneous population in a high-rise building and consider related issues, such as evacuation strategies and evacuation under various situations. This dissertation consists of three subtopics, as follows. First, we identify how seriously residents with disabilities affect the evacuation of other residents. To do so, we analyze the trends of the evacuation times in two population scenarios, homogeneous (i.e., only residents without disabilities) and heterogeneous populations (i.e., residents with and those without disabilities), while we vary the size of populations and the compositions of disability types of residents. According to our experimental results, residents with disabilities significantly delay the evacuation process by causing congestion and blocking phenomenon. We also present regression models to help evacuation administrators ensure the safe evacuation of all residents, by controlling the number of residents and evacuating residents with disabilities efficiently. Second, we present new evacuation strategies for heterogeneous populations, and compare them with a simultaneous evacuation strategy. From the experimental results, we find that a vertically phased evacuation strategy that varies delay times by the physical location is not useful for the simulated building, but a horizontally phased evacuation strategy that applies a fixed evacuation delay to residents with wheelchairs reduces the overall evacuation time. However, because delaying evacuations of a specific group of individuals may not be ethical or accepted, we test an evacuation strategy that allows residents with wheelchairs to use elevators, and the experimental results show the effectiveness of the strategy. In addition to the evacuation strategy considering people with disabilities, we propose an evacuation strategy based on the flow of residents, using the network flow model. We obtained better results in terms of the evacuation time using the proposed approach, and found that it is more applicable than phased evacuation strategies. In addition, the proposed approach is not limited to the simulated building
it can be generalized to other buildings that consist of different structures. Finally, we discuss the evacuation of heterogeneous populations under semi-panic situations, which we distinguish from normal evacuations and panic situations. Before exploring the effect of evacuation under a semi-panic situation, we extend a previous evacuation model that we used and identify the effect of fatigue (which has rarely been presented in studies previously) and the various speeds for each individual, instead of fixed speeds for each type of individual. Considering these changes in residents’ behavior, we not only identify the effects of each change but also aggregate them (i.e., mental disorientation, faster walking speed, variance in walking speed, fatigue accumulation), and, through which, we define the “semi-panic evacuation.” In a semi-panic evacuation, behaviors and features of residents are different from those in non-panic evacuations. The findings from this dissertation will not only provide useful information for the practitioners, such as evacuation administrators who want to develop new evacuation strategies or control the policy of the building, but also may address the effects of several factors, including heterogeneous populations, new evacuation strategies, the fatigue effect, and the semi-panic situation in evacuations of high-rise buildings.
현대사회에서 재난상황과 같은 상황을 대비한 고층빌딩대피는 매우 중요한 문제이다. 따라서 많이 연구자들에 의해서 연구되고 있고 실제 대피현상과 비슷하게 반영하기 위해서 많은 노력을 기울이고 있다. 그러나 대부분의 대피연구들은 인구의 10 ~ 20% 정도를 차지하고 있는 장애인들을 고려하지 않고 있다. 본 연구에서는 장애인을 포함한 고층빌딩대피와 관련된 3 가지의 문제를 다룬다. 첫째로 장애인을 포함하였을 때 어떻게 차이가 나는지를 확인하였다. 이를 위해서 장애인을 포함하였을 때(heterogeneous population)와 포함하지 않았을 때(homogeneous population)의 대피시간을 비교해보고 이 차이가 인구수나 구성비에 따라서 어떻게 변화하는지 확인하였다. 그 결과, 장애인들에 의한 혼잡현상이나 이들이 다른 사람들의 진행을 막는 현상으로 인해 전체 대피를 느리게 하는 것을 확인할 수 있었다. 그리고 회귀분석모델을 통해서 대피관리자들이 빌딩의 수용한계를 넘지 못하게 조절할 수 있도록 적정한 인원을 도출하는 방법을 제안하였다. 두번째로 장애인을 포함하였을 때 효율적인 대피전략을 제안하고 한번에 모두 탈출하는 전략(simultaneous evacuation)과 비교해보았다. 실험결과에 따르면 층별로 탈출을 달리하는 전략(vertically phased evacuation)은 실험대상건물에 유용하지 않았지만 사람의 유형에 따라 출발을 달리하는 전략(horizontally phased evacuation)은 simultaneous evacuation 보다 효율적인 것을 확인할 수 있었다. 그러나 이 방법과 같이 사람의 유형에 따라 출발을 달리하는 것은 형평성의 문제가 있기 때문에 실제로 적용하기는 어려운 방법이다. 그래서 이 사람들을 엘리베이터를 이용하여 탈출시키는 전략을 실험해보았고 효과적인 것을 확인할 수 있었다. 그리고 좀 더 효율적이고 현실적인 방법을 제안하기 위해서 network flow model 에 기반한 대피전략을 구하는 방법을 제안하였다. 이 방법으로 통해서 simultaneous evacuation 보다 효율적인 대피를 할 수 있음을 보였고 이 방법은 phased evacuation 보다 좀 더 현실적인 방법이 될 수 있다. 그리고 이 방법은 실험건물에만 국한된 전략이 아닌 건물의 레이아웃이나 구조가 다르다고 하더라도 쉽게 적용할 수 있다. 마지막으로 일반적인 대피훈련상황이나 panic 상황과는 다른 semi-panic 상황에서의 대피에 대해서 논의를 진행하였다. 이 상황의 대피에 대해서 연구하기 전에 기존의 연구모델을 확장하여 기존 대피연구에서 전혀 고려하고 있지 않은 피로도에 따른 속도감소와 이동속도에서 분산에 따른 효과를 고려하고 효과에 대해서 확인해보았다. 이러한 효과와 더불어 일반적인 대피상황보다는 좀 더 다급한 상황에서 일어나는 변화 (혼란, 좀 더 빠르게 걷기, 피로 누적 등)를 semi-panic 상황에서 사람들에게 일어날 수 있는 변화로 가정하고 그 효과에 대해서 확인해보았다.
Article Type
Files in This Item:
There are no files associated with this item.


  • mendeley

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

Views & Downloads