Evacuation Simulation Using Cellular Automata Model With Forces Essentials

Long before the history of human, fire has become one of the most permanent and severe disasters. Besides great losses of personnel and property, fire also endangers the fields such as energy, resource and environment. Fire has been getting increased along with the development of national economy. Safety problems are getting more attentions that ever before. How to prevent and reduce the injuries and deaths has been the most important thing on the importance list of public safety. Exploration and discovery of typical characters and fundamental rules of pedestrian evacuation, and design guidance of buildings and public places are main methods in fire disaster prevention.There exist interactions among pedestrians and between pedestrian and environment in evacuation. These interactions including attraction, repulsion and friction play key roles in human evacuation behaviors, speed and efficiency. Social force model, a multi-particle self-driven model introduced in recent years can represent those three forces but with low simulation efficiency because it is a continuous model with complex rules. Discrete models, which have been studied widely, such as cellular automata model and lattice gas model, have simple rules and high simulation efficiency. But their simulation results are far from preciseness and reliability before influence of interactions above is well modeled. Aiming at that, here we introduce a new cellular automata model called CAFE based on traditional models, in which repulsion and friction are modeled quantitatively. It is indicated that results from CAFE model consist well with those from social force model on many aspects such as human behaviors, evacuation time and the "faster-is-slower" phenomenon, but with high efficiency.For the purpose of quantification of pedestrian's view field range, we gave in-depth and meticulous analysis of the classic phenomenon—bidirectional movement, in the intention as the expansion and precization of CAFE model. We introduced two CA models: one with extended interaction radius, the other with the traffic rule breaking effect. And it's found that: the critical entrance density does not depend on system size; parameters such as critical blockage entrance density, movement velocity and its change trend, channel occupancy, demonstrate an obvious change comparing with the classical model; pedestrian movement efficiency rises. Research on counter flow model considering traffic rule breaking shows that: the default traffic rule is not always positive to the counter flow in all situations. It depends on the game result of these two opposite sides: to use the channel width as much as possible and to avoid the interference from the other group as far as possible.Then based on the research work, we developed a pedestrian evacuation evaluation software named SafeGo (Software Copyright No. 049051) for the performance-based assessment of buildings. And it's been used widely for buildings' fire protection safety performance design on many applications, such as Olympic Stadium, the Potala Palace, QinShan nuclear power plant, airport terminal buildings, shopping malls and storehouses.