Experiment and Computer Simulation of Human Behavior during Earthquake Induced Fire

为了收集人类在地震火灾过程中的行为数据，对40名受试者进行了两种不同条件下的迷宫实验。还提出了一种新的计算机模拟模型来分析人口密集的大型地下设施中的疏散问题。主要发现：从迷宫实验中发现，受试者一旦失去参考位置，方向感的恢复需要时间。受试者的行为很大程度上受亮度和烟雾的影响。将疏散行为分为三种类型，提出了疏散路径分维作为评价疏散行为复杂性的新指标。利用迷宫实验和实际火灾的数据，检验了分维在疏散路径分类中的应用。通过研究发现，分维是量化疏散行为复杂性的一个很好的指标。提出了一种基于势模型的计算机模拟方法。以前的模型不能有效地处理道路上的障碍。为了验证该方法对人体行为分析的充分性，将该方法应用于一个简单的走廊模型和一个真实的地铁站。通过将计算机模拟结果与以往对人类行为的研究结果进行比较，发现该方法是模拟人口密集的地下设施中人类行为的一种良好而实用的方法。

To collect data on human behavior during earthquake induced fire, a maze experiment with two different conditions was performed using 40 subjects. A new computer simulation model is also proposed to analyze evacuation in large and populous underground facilities. The major findings are :From the maze experiment, it was found that once a subject has lost reference positions, it takes time to recover the sense of direction. The behavior of a subject was greatly affected by brightness and smoke. The behaviors can be categorized into three patterns.The fractal dimension of evacuation path is proposed as a new index to classify the complexity of evacuation behavior. The data from the maze experiment and from an actual fire were employed to examine the use of the fractal dimension in classifying evacuation paths. From these investigations, the fractal dimension is found to be a good index to quantify the complexity of evacuation behavior.A new computer simulation method based on potential model is proposed. Previous models cannot efficiently handle obstacles in the path. To check the adequacy of the method for human behavior analyzes, this method is applied to a simple corridor model and a real subway station. By comparing the results of the computer simulation to the results of past research on human behavior, the proposed method is found to be a good and practical method to simulate human behavior in populous underground facilities.