摘要

地铁结构的复杂性和人员行为的随机性决定了在应急状态下地铁人员疏散过程中存在众多的不确定。人员安全疏散时间和每个出口的人流率是评价疏散结果的重要指标，在计算疏散时间和出口人流率过程中涉及到很多不确定因素，往往都是直接忽略不确定性或者采用保守的安全系数法进行计算，缺乏量化这种不确定性的方法，使得疏散计算结果与实际结果存在偏差。有效的获取不确定性参数并加以量化是地铁疏散系统设计与评估中值得深入探讨的问题。

通过查阅参考文献，整理了地铁车站人员疏散相关的理论知识，从地铁车站建筑环境、人员疏散行为特点和突发事故环境特点三个方面对地铁车站人员疏散的影响因素进行了分析，并对地铁疏散的安全疏散时间进行了分析，阐明了相关规范的要求，为建立仿真模型奠定了理论基础，为本文人员疏散结果分析建立了分析标准。

对地铁疏散不确定性从模糊不确定性、随机不确定性、空间不确定性和时间不确定性四个方面进行分析。通过对目前常用的不确定性分析方法进行分析比较，为本研究不确定方法的选取奠定了基础。根据地铁车站人员疏散的特点，建立了本文的不确定分析体系，为本文针对实际案例进行不确定分析提供了理论支撑。

根据不确定理论，以某地铁车站为背景对人员疏散的不确定性进行了研究。在地铁疏散中的不确定性参数及其概率分布的基础上，基于拉丁超立方抽样的方法获得大量的输入样本，并通过Monte Carlo仿真技术模拟生成并计算输出结果，从而通过散点图、概率密度直方图等统计学的方法初步判定因素之间的敏感性。基于单因素敏感性分析方法，定量分析不确定因素对疏散结果的影响。通过比较疏散时间根据影响因素变化波动大小，判别因素的敏感程度，最终得出对该地铁车站疏散安全状态的最敏感的因素，及整个地铁车站的安全疏散可靠度，为地铁车站运营管理及人员安全疏散提供指导建议。

关键词：地铁疏散；不确定性分析；敏感性分析；Pathfinder仿真；Monte Carlo模拟

Abstract

The complexity of subwaystructure and the randomness of the conduct of members determine the uncertain duringthe subway evacuation in the state of emergency. As the important indicator ofevaluating egress results, the safety egress time and the passenger flow ofeach outlet contains uncertainties. Uncertainties were often ignored orcalculated by conservative safety factor method. The lack of quantitativemethods of uncertainties makes deviation of calculated results and the actualresults of egress.Effectively accessing and quantifying parameter uncertainties during subwayegress system design and evaluation deserve further exploration.

The article analyzes the subwaystation occupant egress from the subway station built environment, evacuationbehavior characteristics and the incidents characteristics three aspects basedon finishing the occupant egress theory. And the safety egress time is analyzedin subway ergess and the requirements of the relevant specifications are clarified,which the theoretical basis for building the simulation model and establishedanalytical standards for the analysis results of occupant ergsee in thisarticle.

The article analyzes the subwayegress uncertainties from the vague uncertainty, the random uncertainty,spatial uncertainty and timing uncertainty four aspects. The analysis and comparisionof the uncertainty analysis methods currently laid foundation for selecting theuncertainty analysis method in this article. According to the characteristicsof subway station occupant egress, the paper established the system of uncertaintyanalysis, which provided a theoretical support for the uncertainty analysis of actualcase.

According to the uncertainty theory,this article studied the uncertainty of occupant ergess based on a metrostation. Based on the uncertainty parameters and probability distribution insubway egress, a large number of input samples are obtained by the latinhypercube sampling method. Then the outputs are calculated by the Monte Carlosimulation, which contributs to preliminary determin sensitivity betweenfactors by some statistical methods. The article quantitative analyzes theimpact of uncertainties on the results of the safety egress based on the singlefactor sensitivity analysis method. The factors’ sensitivity are discriminatedby comparing the fluctuations range of ergress time. At last, the analysisresult shows the most sensitive factor to the subway station ergress during securitystate and the reliability of the whole station safety egress, which provides instructionfor the subway station operation management and occupant safety egress.

Key words:Metro Egress; UncertaintyAnalysis; Sensitivity Analysis; Pathfinder simulation; Monte Carlo Simulation