基于元胞自动机的火灾场景行人流疏散仿真研究

Evacuation Simulation in Narrow Passage Under Fire Scenario Based on Cellular Automaton

摘要:

基于元胞自动机的场强模型在二维平面行人流疏散问题的研究中已得到了广泛应用.已有模型主要描述行人基于出口位置并跟随其余行人进行疏散的行为特征，未充分考虑火灾蔓延和局部拥堵对行为选择的影响，难以准确模拟这些情景下的疏散过程.实际情况下，火灾环境将显著影响行人的移动方向，周边行人流的拥堵程度则影响行人的移动需求.提出基于火灾场景的移动方向，考虑火灾导致的恐慌对移动方向的影响；基于行人流场景的期望速度，考虑行人流中局部拥堵对移动需求的影响，对已有场强模型进行修正.参数分析表明：考虑基于场景的移动方向和期望速度规则后，疏散时间和平均死亡人数均呈现变化；不考虑移动方向和期望速度的行为特征将低估疏散总时间，并高估火灾导致的平均死亡人数.

Abstract:

The cellular automaton (CA)based floor field model has been widely applied to addressing the issues concerning the 2D evacuation. It is found that the existing floor field models mainly focus on the behavioral features based on the location of the exit and following others in the lead, but the influence of fire expansion and local congestion is not fully considered, resulting in the inability to precisely simulate the evacuation process in such scenarios. In reality, the moving directions of occupants are heavily dependent on the expansion of fire, while the moving needs of occupants are remarkably influenced by the congestion in the adjacent pedestrian flow. Correspondingly, in this paper, the fire scenario based moving direction is proposed to describe the moving direction affected by fireinduced panic; the pedestrian flow based expected velocity is proposed to address the moving need affected by local congestions, in a bid to modify the present models. The parameter analysis based on the proposed model indicates that the overall evacuation time and the average death toll vary when considering the scenario based moving intention and expected velocity; the overall evacuation time will be underestimated while the average death toll will be overestimated if these two behavior features are not incorporated.