برآورد تاثیر ساکنان معلول در تخلیه ساختمان های بلند: یک مطالعه شبیه سازی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|10092||2012||13 صفحه PDF||سفارش دهید||9004 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Simulation Modelling Practice and Theory, , Volume 24, May 2012, Pages 71-83
In this paper, we study how seriously residents with disabilities affect the evacuation of other residents in a 24-story high-rise building environment through an agent-based simulation model. In particular, we analyze the trends of the evacuation times in two population scenarios, homogeneous (i.e., only residents without disabilities) and heterogeneous residents (i.e., residents with and 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. Our experimental results also indicate that the differences in the evacuation time of the homogenous and the heterogeneous population scenarios become more significant as the population size increases because of serious congestion from the increased population size and blocking from the increased proportion of the handicapped. Finally, we present regression models and controlled evacuation strategies that help evacuation administrators ensure the safe evacuation of all the residents by controlling the number of residents and evacuate residents with disabilities efficiently.
Efficient evacuation plans are strongly required in large and complex structures such as high-rise buildings, airports, schools, stadiums, and large vehicles including airplanes and ships to minimize the casualty of people during catastrophic events. In particular, there are strong needs of developing evacuation plans that consider residents with disabilities as well as residents without disabilities due to the significant proportion of residents with disabilities in our community. For example, about 12.1% of population in the United States has been known to suffer from physical, mental, or emotional disability  while between 14% and 17% of the UK population can be classified as the disabled . While these residents with disabilities are more likely to suffer from natural and human caused disasters and slow down the evacuation of other residents due to their larger space requirements and slower evacuation speeds, only very few studies consider them explicitly. In addition, there is a significant lack of studies on the relationship between the structural characteristics of the building and the evacuation of residents with disabilities even if Sagun et al.  mentioned that one of the most important contributions of evacuation simulation is improving the design of the built environment. While several studies , , , ,  and  considered heterogeneous population including residents with disabilities, their heterogeneity designs only consider the very limited characteristics of residents without disabilities such as age, gender, movement speed, and physical size based on the characteristics of residents without disabilities. Christensen et al.  also indicated that many evacuation models do not address residents with disabilities appropriately. We conjecture that the evacuation time of heterogeneous residents with and without disabilities in a building or a large vehicle can be significantly different from that of only residents without disabilities in the same building. Therefore, evacuation administrators need to develop a comprehensive evacuation plan that explicitly considers residents with disabilities because of their unique physical and mental needs under emergency situations. For example, most residents with disabilities such as wheelchair users cannot evacuate using stairs easily. They are also typically slower than the non-disabled and hence may delay the evacuation of other residents by blocking others in narrow paths such as stairs. However, most of evacuation planners and administrators do not know how far residents with disabilities delay the evacuation of other residents and hence cannot revise evacuation plans accordingly. Therefore, in this paper, we intend to estimate the impact of residents with disabilities on the evacuation time of other residents in a 24-story high-rise building environment through an agent-based simulation model. For this purpose, we simulate and analyze the trends of the evacuation times of two population scenarios, homogeneous (i.e., only residents without disabilities) and heterogeneous residents scenario (i.e., residents with and without disabilities), while we vary the size of resident populations and the resident compositions of various disability types in the population. Another research question we like to explore in this paper is to identify several places in the building that are most likely to cause congestion and reduce the number of evacuees who can exit through the bottlenecked place simultaneously, resulting in so called “capacity drop” . Once several such places are identified, evacuation administrators pay special attention to improve the structural design or clear out any obstacles around such places. Finally, we like to develop regression models based on simulated data sets to help evacuation administrators determine the maximum size of resident population whom can be evacuated within a desired evacuation time or help them estimate the time needed to evacuate all the residents with and without disabilities. This information from regression models then can be used to ensure the safe evacuation of all the residents by regulating the number of residents and visitors in the building. In addition, we also like to develop new evacuation strategies that control the evacuation process of residents by either evacuating residents at lower levels first or evacuating residents at higher levels first to minimize the possibility of congestions and capacity drop phenomenon. Practically, we implement these strategies by controlling the propagation of emergency alarm so that residents only on the intended floors can hear it and start to evacuate. Note that residents with disabilities have different physical and mental characteristics and needs. Therefore, three different evacuation strategies for residents with disabilities have been studied . The first strategy is to use the designated refuge areas in a building. This strategy is based on the observation that wheelchair users cannot evacuate by themselves via steep stairs during an emergency, and hence it is better for them to move to the designated refuge area where they can wait for rescuers. The second strategy being considered is to provide safe elevators that can be used for emergency evacuation by residents with disabilities in some special buildings. The last strategy is to develop specific evacuation procedures for residents with disabilities with the help of healthy residents. The “buddy” system, for example, identifies one or a few persons who have the responsibility of looking after individuals with disabilities in case of an emergency. In the buddy system, the residents with disabilities will be carried out by buddies with or without using special devices. In this study, we take a buddy system approach to evacuate residents with disabilities. This paper is organized as follows. We first briefly review the related literature in Section 2. Section 3 presents the agent-based simulation model, the simulated building, and the resident configurations for this study. In Section 4, we present our findings with regard to the effects of resident compositions and sizes, and structural characteristics of the building that are most likely to create congestions on the evacuation times. In Section 5, we develop and compare two controlled evacuation strategies that evacuation administrators can use to efficiently evacuate all the residents while minimizing the evacuation times. Finally, Section 6 provides the conclusion of the paper and suggests several directions of further research.
نتیجه گیری انگلیسی
In this study, an agent-based simulation system is adopted to study how the compositions and sizes of resident population affect the overall evacuation times of other residents in a 24-story high-rise building. In our experiment, we control the compositions of the population by considering two different population scenarios, PopWoD in which all the residents do not have any type of disabilities and PopWD in which residents with and without disabilities co-exist based on US Census Bureau Data. For each population scenario, we also vary the sizes of population generating six different population sizes (i.e., 107, 213, 426, 852, 1704, and 3408 residents). Our experimental results reveal that there exists a positive relationship between the population size and the evacuation time in PopWoD scenario mainly because of congestion effects from more residents. While the same positive relationship between the population size and the evacuation time is also discovered in PopWD, the relationship turned out to be much stronger. We attribute this finding to another factor, blocking effect mainly caused by residents who require more personal spaces and who evacuate slowly (e.g., residents on wheelchairs). Note that the blocking effect occurs even if the number of residents with disabilities is minimal. In particular, if the narrow stair area is only way of evacuation, even residents with faster mobility have no other options but to follow slower residents ahead because there are no spaces for them to pass, resulting in significant increases in evacuation time. We also find that when serious congestion occurs due to the increase of population size and/or blocking effect from residents with disability, capacity drop phenomenon quantified by measuring the flow rate and density of each floor can also occur. When capacity drop phenomenon occurs, residents cannot evacuate at the maximum flow rate of the facilities along the evacuation routes. Our simulation tool can be very useful for evacuation administrators or evacuation planners to develop new evacuation strategies considering unique characteristics of residents in their building. For example, we develop the regression models for both PopWoD and PopWD population scenarios to identify a relationship between population sizes and mean evacuation times using simulation data. These regression models can be very useful for evacuation administrators to estimate the maximum number of residents whom can be evacuated within a certain evacuation time or to estimate the evacuation time that would take to evacuate all the current residents in the building. They can also use such information to control the number of residents and visitors in the building, develop more efficient evacuation routes for residents with disabilities, or remove and renovate structural obstacles of the building. In addition, we develop and compare two evacuation strategies that control the evacuation process of residents by either evacuating residents at lower levels first or evacuate residents at higher levels first. Practically, we implement these strategies to control the propagation of emergency alarm so that only residents on the intended floors can hear it and start to evacuate. Our experimental results show that controlled evacuation strategies that evacuate residents on lower floors first are statistically significantly superior to the strategies that evacuate residents on higher floors first mainly because evacuating residents at lower floors reduces the severity of congestions and capacity drop phenomena. One of possible extensions of the current study is to develop optimal evacuation strategies that consider the structural characteristics of the building such as stair areas that are likely to cause congestions, the length and space of the hall area, and evacuation elevators. For example, we are currently investigating how to use the information of congested floors and stair areas (3rd floor–10th floor) during the simulated evacuation to develop a new controlled evacuation strategy and how much control delay should be applied to minimize the evacuation times of all the residents. In addition, we are also considering various times of control delay for our controlled evacuation strategies depending on the types of residents. For example, we like to find out an optimal control delay that is sufficient for residents without disability to evacuate first to avoid the blocking effect caused by residents with disabilities while it does not jeopardize the survival opportunity of residents with disabilities. In particular, evacuation strategies that allow residents with disabilities to use an emergency elevator can be very useful since it is almost impossible for helpers to carry them down stairs from 20 or higher floors without risking their lives. Along this line of research, we will investigate who can use it (e.g., all the residents with disabilities only vs. all the residents with disabilities and those who can help them to get on and off it), how to operate it (e.g., should it stop at every floor, every even (or odd) floors, or every 5 floors from the top), and how to combine this evacuation strategy with other traditional ways of evacuating residents with disabilities (e.g., buddy system).