Experimental Study On Evacuation Behaviours

At present, simulation is the major way to study evacuation. However, due to lack of evacuation data, the verification of evacuation models is still a challenging problem which has not got a good solution. This problem is particularly severe in China. Due to the lack of evacuation data for Chinese population, domestic simulation studies on evacuation often rely on foreign data. However, evacuation behaviours have strong dependence on human characteristics, such as physical feature, cultural backgrounds, habits and emergency training. So the evacuation behaviours of Chinese population must have some differences with foreign population. Using foreign data to study evacuation of Chinese crowd will inevitably greatly reduce the accuracy and credibility of computer simulation. Therefore, it is necessary to conduct research work on data collection and analysis of evacuation behaviour of Chinese population. In this study, data collection and analysis of evacuation behaviour of Chinese population is carried out by means of questionnaire survey, evacuation drill and the observation of pedestrian flow in a subway station.First of all, a questionnaire survey is conducted to make a qualitative analysis of evacuation behaviours. Taking fire incidents for instance, some typical evacuation behaviours during disasters are statistically analyzed. The impact of human characteristics, such as gender, age, education level and training, on evacuation behaviours is studied.Then, an evacuation drill is carried out to make a quantitative study of evacuation behaviours. Comprehensive methods including questionnaire survey, empirical formula calculation and computer simulation, are used to make a systematic analysis of the drill experiment. Route selection of individuals, evacuation time, and crowd movements at exits and on stairs are studied. The relationship between exit flow and width is studied. It is found that the unit flow of exits is not a constant and has a close tie with crowd density and evacuation order. The formulas of the relation between unit flow, velocity and density on stairwells are summarized by drill data.Meanwhile, drill data are used to verify and calibrate the parameters of empirical formulas and computer model. Togawa formula is used to calculate evacuation time. The formula's parameters, movement speed and unit flow, are calibrated by the drill data. Two kinds of fine network evacuation models are used to make a simulation analysis on the drill. The comparison on simulation accuracy between the two models is made with the same input parameters. It is proved that fine network model is capable of simulating the process of an orderly evacuation, in which the physical interactions between individuals at bottlenecks are not serious. Simulations of the drill with different input values of pedestrian speed are conducted. It is found that the accuracy of simulations have a great dependence on the model's pedestrian speed. Factors that influence the pedestrian speed are discussed.Finally, the observation of pedestrian flow in a subway station is carried out to make a comparison with pedestrian flow in evacuation drill. The characteristics of crowd evacuation movement on stairwells at normal state and under emergency are compared. The results show that the movement speed and unit flow of the pedestrians in the drill is greatly higher than that in the subway station at the same crowd density.In this paper, large amounts of data of evacuation behaviours are collected by experiment and investigation. Empirical formulas and evacuation models are verified and calibrated. The result of this study can provide reference for evacuation study and performance-based design.