【作者】 李健；

【导师】 杨立中；

【作者基本信息】 中国科学技术大学 ， 安全技术及工程， 2008， 博士

【摘要】 人员疏散是一种多参数影响的复杂过程,需要研究的内容也非常广泛,其中就包括人员运动规律和他们对环境信息行为反应的研究。由于从疏散演习和灾后调查获得的数据比较缺乏,所以目前这个领域的研究以计算机模拟为主。本博士论文的第一章,作者在对人员疏散研究内容和方法综合分析的基础上确立了本文的主要目的:试图建立合理的,充分考虑人员特性行为和环境信息影响的紧急情况下人员疏散模型,并通过该模型对人员运动的特殊现象进行研究,以期了解人员运动规律和影响疏散效率的关键因素;另外,创造条件进行疏散实验,通过第一手实验数据验证模型的可靠度和适用度,同时为改进模型提供思路;最终希望为建筑物的设计和疏散演习方案的策划提供理论支持。在接下来的两章中,作者提出了一种考虑了个体特性的人员运动元胞自动机(CA)模型,并借此研究了正常情况下人员自身的运动规律。其中,第二章主要关注了以“推挤-穿行”为代表的非常规运动行为对行人流动力学过程的影响。我们提出了一种描述个体行为的数学方法,特别是“推挤-穿行”行为。通过对相向人员流和二维人员流的模拟,定量分析了个体的非常规移动对宏观运动的影响,着重考察了动力学转变过程。我们发现,基于不同的人员密度和微观特性,人员动力流过程可分成4个相位:自由相、过渡相、受限相和阻塞相。另外,对循环边界条件而言,相变临界密度与系统尺寸无关。我们调查了个体移动路线和宏观相位的关系。最后,本章比照中国人口实际尺寸标准,指出了“推挤-穿行”行为可接受的人员密度范围。在第三章中,我们重点考察了以“右行偏好”(人们偏向于沿道路右侧行走)为代表的非对称选择心理对人员运动的影响。本章用CA模型模拟了通道中人员相向流。我们提出了“右偏强度”κ,即右移概率与左移概率的比值。模拟中再现了从自由流到堵塞相的动力学转变过程。采用循环边界条件时,对于确定的κ,相变临界密度与通道尺寸无关。参照国际上心理学和社会学的研究成果,本章考察了κ=1,2,8的情形,发现κ=8时最符合实际情况。在此基础上,我们讨论了“右偏强度”的计算表达式。在本章最后,我们将模拟结果与日本学者的实验结果(M.Isobe,2004)做了比较,证实当人员密度低于危险值时,偏右行走能提高通道使用效率。另外,我们还探讨了通道中行人流的踩踏现象。在对正常情况下人员运动规律有一定了解后,我们把注意力集中在如何描述紧急疏散时人员的行为和反应。这也是本文的核心研究部分。在第四章中我们提出了一种考虑环境信息影响的多速度人员疏散模型。作为人员疏散过程中必不可少的因素,“静态信息”(建筑结构、空间惯性等等)和“动态信息”(警报信号等等)在本章中通过模拟和实验得到较为深入的研究。我们展示了7种情况(通过控制疏散人数、能见度、警报、出口宽度等条件)下教室疏散的实验和模拟结果,主要关注了个体逃生路线、时间和人员流率等。尽管人员运动有较强的随机性,但我们的模型仍可以很好地半定量地再现实验观察结果。在第四章术,我们规划了基于信息传播理论的人员疏散模型框架,主要提出了紧急疏散中狭义和广义信息的概念。它可以作为从理论方法到应用软件的一种技术途径。最后,我们在第五章中对目前工作做了小结和展望。更多还原

【Abstract】 Occupant evacuation is a kind of complex process with multi-elements, including the behavioral response to environment information and people’s movement. Research work on this field is mainly developed by computer simulation, because the available data from realistic evacuation drills is scarce.In Chapter 1, methods and contents in studying occupant evacuation are introduced. The key idea of this paper is also summarized: to develop an occupant evacuation model involving the individual characteristics and the impact of enviornment information （such as: the sound of fire alarm）; to study some special phenomena observed in evacuation and to find out some factors that could have an impact on evacuation efficiency by using this model. In addition, the simulational results should be compared with experimental results to examine the model’s applicability and improve it.In the two following chapters, a occupant movement model based on Cellular Automaton （CA） has been introduced. Some individual characteristics have been involved and studied within normal condition. In Chapter 2, the dynamics process of movement with "jostle- through" behavior has been focused. A numerical method is proposed to describe individual behaviors, including "jostle-through": P_ex is introduced as the success probability of a single winning through. Counter flow and 2-dimension flow have been studied by means of simulation. We find that there are four phases of the dynamics flow, i.e., free phase, transition phase, restriction phase and jam phase, dependent on the system density and the micro-characters. For circular boundary condition, it is also found that the critical density is independent with the system size. Individual route is investigated, and the practical interval of pedestrian density for jostles is given, which is according with the human dimensions of Chinese population.In Chapter 3, the right-side-preference （people prefer to walk on the right-hand side of road or channel） has been investigated, as an asymmetrical behavior that should not be ignored in occupant movement. In this chapter, pedestrian counter-flow in a channel is simulated using CA model. The right-preference intensity, k, is introduced, defined as the ratio of the right-moving probability to left-moving probability. In simulations, the dynamical transition between fluid and jammed phase is presented. With a fixed k, the critical density is independent of the channel size. According to research results on physiology and sociology, k =1,2,8 have been tested, and k =8 is satisfied in this work. Furthermore, we have discussed one updated equation for k. At the end of this chapter, simulation results are compared with Japanese experiments （M.Isobe, 2004）. It is found that right-preference is effective when the density is below critical. The model is shown to be useful to simulate and analyze this situation numerically.Based on people’s movement under normal condition, we focus our attention on how to describe the occupant behavior and response in emergency. It is the most important part of this paper. In Chapter 4, an evacuation model involving environment information with multi-velocity occupant has been proposed. As essential elements in occupant evacuation, the "static informaiton"（building structure, spacial inertia,etc.） and "dynamic information" （sounds of fire alarm, etc.） have been studied by means of simulation and experiments. We have presented experimental and simulation results on the evacuation of a classroom in 7 conditions （occupant number, visual-field, alarm, exit width, etc.）, focusing on the individual escape times and the escape flow rates as a function of time. Despite the stochastic nature of occupant movement, the empirical observations could be semi-quantitatively reproduced by our model.At the end of Chapter 4, an evacuation model based on information transfer system is planned, as an extensive development of the theoretical method into practical software. The special and general theory of information transfer within evacuation is introduced in this chapter.In Chapter 5, the main work of this paper is summarized and their development in the future is expected.更多还原