非常规突发情况下大规模人群疏散的不确定性研究

发布时间：2018-04-20 19:18

本文选题：非常规突发灾害 + 大规模人群疏散　；参考：《中国科学技术大学》2013年博士论文

【摘要】：近年来,在全世界范围内具有灾难性后果的非常规突发事件时有发生,大规模群体疏散问题日益得到公共安全领域的高度重视。迄今为止,对于大规模人群疏散的微观或宏观模型研究已有很多,有力推动了非常规突发事件应急处置科技领域的发展。但是,关于非常规突发事件下大规模人群疏散的不确定性尚没有得到充分的认识。为了揭示非常规突发事件下大规模人群在疏散过程中所受到的诸多不确定因素的影响和作用模式,我们采用理论分析和数学建模相结合的方法,以非常规突发灾害为背景,从突发灾害、外部疏散引导以及大规模疏散人群这三个层面,研究影响大规模人群疏散的随机不确定性,从效率和风险的角度,为突发灾害应急疏散提供理论和技术支持。在非常规突发灾害层面,首先对非常规灾害与常规灾害的随机性分析与预测进行了一些探讨。以常规灾害中的城市火灾为例,建立了基于城市火灾幂律分布特征的重特大火灾事故发生概率预测模型,而非常规灾害在时间和空间尺度上均具有更多的不确定性和复杂性。因此,涉及到非常规灾害相关参数的设置和分析时,主要以人为假定的形式进行。在对灾害扩散模式及危害程度进行合理假设的基础上,运用动态网络流方法,建立了一个综合考虑疏散优先级和灾害扩散一般模式的多源多汇(MSMD)大规模疏散模型,并运用CCRP算法进行了疏散规划的求解及对比。进一步提出并建立了“路径通行效率风险”(RTE风险)的概念及其定量评估框架模型,基于该模型获得的各路段在灾害发生后不同时刻的ITE风险值,可有效反映出疏散路网在灾害环境下,综合考虑了效率和风险之后“适合通行”程度的动态变化。在疏散引导层面,一是研究了大规模人群疏散时极易出现的恐慌情绪的传播特征及其对疏散的影响。运用系统动力学方法,构建了一个大规模人群疏散的定性仿真模型,通过对模型实施各种不同的输入方案,发现了有无疏散引导时由于灾害氛围加剧而出现的恐慌受控或失控状况的变化规律,以及灾害氛围下人群中恐慌情绪的蔓延更多地是受到占主导地位情绪的作用。该定性仿真模型很好地揭示了疏散系统中关键要素之间的相互作用关系以及影响疏散中恐慌情绪传播的不确定性因素,并且再现了疏散出口处“快即是慢”的典型现象。二是结合疏散引导对大规模疏散风险进行了量化研究。提出了大规模人群流动特征密度的概念,并基于连续人群流动理论和相关经验公式,推导获得了影响大规模人群流动的一系列特征密度。基于所获得的特征密度,结合排队理论,建立了一个无限人流过桥模型,并通过计算相关的系统效率指标,获得了不同密度人群在不同疏散引导策略下的疏散效率变化特征。基于理论分析和排队模拟所获得的五个特征密度,提出了一个在大规模疏散中判断疏散策略能否提高疏散效率的“人群密度风险轴”,以人群密度风险轴上的三个区间代表大规模疏散时针对疏散策略有效性而言的三种不同特征的人流,即有效流、临界流和无效流,并通过疏散风险的合理数值匹配,对这个人群密度风险轴进行了定量诠释。在大规模疏散人群层面,引入并量化了人群生理心理因素对大规模疏散的影响,建立并修正了大规模人群疏散路径选择的随机Markov模型,并基于Markov过程的概率描述,以CO毒气瞬时泄漏扩散为事故背景,详细分析了疏散人群在多种因素,尤其是生理和心理因素影响下的疏散不确定性。通过分析清空时间在相关参数作用下的变化规律,发现在单因素变化情况下,滞留人数的对数与疏散时间之间存在分段线性特征；清空时间随着初始疏散人数的增加呈现对数线性增加的规律,随着节点最大容量的增加呈现线性减小的规律。此外,徒步疏散情况下人员心理恐慌对疏散速度的修正,整体上对于疏散结果的影响不大,而毒气影响下的人员生理风险对疏散结果影响显著,必须在涉及诸如毒气泄漏等突发事故的大规模人群应急疏散中予以重点关注。
[Abstract]:In recent years, unconventional emergencies with disastrous consequences have occurred in the world, and mass evacuation has been paid more and more attention in the field of public security. So far, there have been a lot of micro or macro models for large-scale crowd evacuation, which has greatly promoted the emergency disposal technology of unconventional emergencies. However, the uncertainty of large-scale crowd evacuation under unconventional emergencies has not yet been fully understood. In order to reveal the effects and modes of many uncertain factors in the process of mass evacuation in the process of unconventional emergencies, we combine theoretical analysis with mathematical modeling. Methods, with the background of unconventional sudden disasters, the random uncertainty affecting the evacuation of large-scale crowd is studied from three levels of sudden disaster, external evacuation guide and mass evacuation, and the theoretical and technical support for emergency evacuation is provided from the perspective of efficiency and risk.
At the level of unconventional emergencies, the random analysis and prediction of unconventional and conventional disasters are discussed. Taking urban fires in conventional disasters as an example, the probability prediction model of heavy fire accidents based on the distribution characteristics of urban fire power law is established, while non conventional disasters are on time and space scales. All of them have more uncertainty and complexity. Therefore, when the parameters related to unconventional disasters are set up and analyzed, they are mainly assumed in the form of human assumption. On the basis of reasonable assumptions about the pattern of disaster diffusion and the degree of damage, a dynamic network flow method is used to build a comprehensive consideration of evacuation priority and disaster expansion. The multi source and multi sink (MSMD) large-scale evacuation model of the scattered general model is used to solve and compare the evacuation planning with the CCRP algorithm. The concept of "path efficiency risk" (RTE risk) and its quantitative evaluation framework model are proposed and the ITE wind of each section obtained at different times after the disaster occurs. The risk value can effectively reflect the dynamic changes of the evacuation network in the disaster environment considering the efficiency and risk.
At the level of evacuation guidance, the first is to study the propagation characteristics of panic and its influence on evacuation. A qualitative simulation model of large crowd evacuation is constructed by system dynamics method. By implementing various different input schemes to the model, it is found that there is no evacuation guide due to the model. The changing laws of panic controlled or out of control, as well as the spread of panic in the crowd under the disaster atmosphere, are more of the dominant emotion. The qualitative simulation model reveals the interrelationship between the key elements in the evacuation system and the influence of the panic mood in the evacuation system. The indeterminacy of the propagation is a typical phenomenon of "fast and slow" in the exit exit. Two is a quantitative study of the risk of large-scale evacuation with evacuation guidance. The concept of mass flow characteristic density is proposed. Based on the continuous crowd flow theory and the related empirical formula, the influence of large scale is derived. A series of characteristic density of population flow. Based on the acquired characteristic density and queuing theory, an infinite pedestrian crossing model is established. By calculating the related system efficiency indexes, the evacuation efficiency characteristics of different density groups under different evacuation guidance strategies are obtained. Based on the theoretical analysis and queuing simulation, the characteristics of the evacuation efficiency are obtained. The five characteristic density, a "crowd density risk axis" which can judge whether the evacuation strategy can improve the evacuation efficiency in large scale evacuation, and the three sections of the population density risk axis, representing three different characteristics of the evacuation strategy, which are the effective flow, the critical flow and the invalid flow, are represented by the large-scale evacuation clockwise. Through the reasonable numerical matching of evacuation risk, the density risk axis of the population is quantitatively interpreted.
At the level of mass evacuation, the influence of the population physiological and psychological factors on the large-scale evacuation is introduced and quantified, and the random Markov model of the mass evacuation route selection is established and amended. Based on the probability description of the Markov process, a variety of factors are analyzed in detail in the background of the transient leakage of CO gas. Especially the evacuation uncertainty under the influence of physiological and psychological factors. By analyzing the variation of the clearance time under the action of relevant parameters, it is found that there is a piecewise linear feature between the logarithm of the number of detained numbers and the evacuation time in the case of single factor change, and the clearance time increases with the logarithmic linear increase with the increase of the number of evacuations at the beginning. The law, with the increase of the maximum capacity of the node, presents a linear decrease with the increase of the maximum capacity of the node. In addition, the correction of the evacuation speed by the psychological panic of the personnel in the case of walking evacuation has little influence on the evacuation results as a whole, and the physiological risk of the personnel under the influence of poison gas has a significant influence on the evacuation results, and it must be involved in such emergencies as gas leakage. Therefore, the large-scale crowd emergency evacuation should be given priority attention.

【学位授予单位】：中国科学技术大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：X93

【参考文献】