Simulation Of Crowd Stampede Based On An Improved Beer Foam Model

With the acceleration of the global urbanization process,the rapid increase of urban population,the frequent holding of large group activities,and the continuous occurrence of various emergencies such as natural disasters,terrorist attacks and group disturbances in recent years,the phenomenon of crowd stampede appears to be complicated and frequent.Crowd stampede has become the main type of accidents which threaten the safety of large public places.The public safety and security of the people meet some serious challenges.Crowd stampede phenomenon is a kind of self-organizing emergence form of large-scale group behavior,which is caused by the rapid accumulation of the panic personnel in the limited space,and its essence is the accumulation of large,complicated,concurrent and non-linear mutual contact forces among the individuals.Because the process of crowd stampede is extremely complicated,the research on crowd stampede is still very challenging.Aiming at the above problems,this thesis carries out the correlation mapping between the crowd trampling force model and the beer foam model.First,this thesis analyzes the complex interaction force among the individual in the stampede,and mainly analyzes the "crowding force" between individual and individual during the stampede.Then,the interaction forces between bubble and bubble and between bubble and solid in the traditional beer foam model are modeled.In order to be as similar as possible to the force model of the crowd,the traditional beer foam model is improved.The spring model is introduced to interconnect the improved beer foam model and the crowd model.Finally,the correlation mapping between the crowd trampling force model and the improved beer foam model is carried out,and the interaction force bubble-bubble and bubble-solids in the beer foam model are mapped into the interaction forces between individual-individual and individual-obstacles in the crowd model respectively.In this thesis,the classical "social force model" is improved,and the individual in the stampede is modeled and collided by the improved "social force model".In view of the phenomenon of "overlap" and "jitter" in the high density population simulation of the improved "social force model",this thesis introduces the collision prediction and the reaction of obstacles to the individual in the "social Force model",which is used to eliminate the problem of "overlap" and "jitter" among individuals.On the basis of the above analysis,an improved beer foam model is used to simulate the crowd stampede events.The experimental results show that the method presented in this thesis can simulate the population of different scale and density,and the collision detection mechanism in the social force model can solve the problem of "overlap" and "jitter" in high density crowd simulation.The simulation experiment in this paper can be applied to the real scene where congestion and stamping accidents have occurred.According to the simulation results,the cause of the accident is found and a reasonable solution is given.It can also be applied to the virtual scene without congestion and stamping accident.The simulation gives a reasonable obstacle layout and control of the flow velocity in the building,which provides a theoretical basis for the evacuation of the crowd in the real scene.