Odelling And Experimental Study Of Evacuation Process Considering Effect Of Fire

Recently, there are more than One hundred thousand fire cases happened in each year, resulting in thousands of casualties. When a fire occurs, whether occupants could egress timely is the most noteworthy problem. In order to reveal the characteristics of pedestrian behavior in fire environment, in this paper, we will study the fundamental human movement behavior by combining experiment and simulation method, and then introduce the effect of fire products on human behavior, finally build an evacuation model coupling with the effects of fire products.We first carried out an evacuation experiment in a teaching building with two exits, and then studied the evacuation process with a multi-grid evacuation model. The basic parameters such as flow, density and velocity of pedestrians in the exit area were measured. The exit-selecting phenomenon in the experiment was analyzed. It was found that pedestrians will select the closer exit even though the other exit is only a little far and empty. In our experiment, over2/3pedestrians used the closer exit. We built a modified biased random walk model to reproduce the experiment. In the model the behavior of selecting closer exit is achieved using the drift direction and the drift force. The former is determined by the shortest distance to the exit, and the latter determines the drift of selecting closer exit or direction. We simulated the evacuation in different scenarios and detected the value of drift force and expired velocity that result in consistent simulation results with the experiment results. Our simulation results afford a calibrated value of the drift force, especially when it is set as0.56, there is good agreement between the simulation results and the experiment results on the number of pedestrians selecting the closer exit, the average velocity through the exits, the cumulative probability distribution of the instantaneous velocity and the fundamental diagram of the flow through exits.In order to explore the microscopic characteristics of pedestrian movement, we then conducted controlled experiments of single-file pedestrian movement, extracted the motion data by using a mean-shift digital image processing algorithm and analyzed the movement characteristics of pedestrians. It was found that the distance headway has a great impact on the step frequency as well as the step length. Considering these special experimental observations, we built a continuous distance model (CDM). In this model, two new insights into the movement algorithm were taken into account. The first one was about the step length. To avoid unreasonable results caused by the simplification of step length in traditional discrete models, a continuous step length was adopted in the model. Moreover, it was observed that pedestrians tend to keep a suitable distance to others to avoid potential collisions, so that two free parameters P1and P2are introduced to represent the proportion of front space that step length accounts for. The other one is about the transition probability. Since the forward transition probability is correlated with the actual step frequency, the dependency between the transition probability and the distance headway is constructed according to the empirical results. We carried out a series of simulation using the CDM model with different numbers of persons in the passageway and different P1and P2values. It is found that when adopting suitable values, especially when P1=0.79and P2=0.62, the flow-density and velocity-density relations produced by the CDM model were in good agreement with the experimental ones. To further verify the model, we analyzed the statistical characteristics of velocities and gave the probability density distribution. It was found that the model presented herein could reproduce further microscopic feature for the scenarios with high density in comparison with the empirical data.It is important and necessary to introduce the effect of fire products on occupants into evacuation models for the reason that exposure to toxic smoke is one of the hazards confronting people in fires. So finally, we introduced the effect of fire products on pedestrian into the multi-grid evacuation model. Some typical fire products, smoke temperature, visibility and carbon monoxide(CO), are emphasized in our model. Their effects on pedestrians mainly consist of three parts:desired movement direction, step frequency of pedestrians and human health. We studied the difference of evacuation routes in normal and fire conditions and also investigated the effects of Heat Release Rate, door width and pre-movement time on the evacuation process with this model. The results reflect the fact that small fire will make egress process faster while large fire will reduce evacuation efficiency significantly. The movement time increases rapidly as the pre-movement increases due to the development of fire. The door width plays a more important role for evacuation in fire than normal condition. We simulated the evacuation process of pedestrians from a hall with two exits. It is found that more and more pedestrians, who move towards one exit at first, will turn back and escape through the other exit as the degree of fire response increases. This leads to the inefficient use of the exit near fire, and increases the evacuation time.The results will be helpful to the understanding of human behaviors, as well as to the building design. Based on the models built in this paper, we developed a software, named FireGo, that may be helpful to fire safety analysis and building design.