“鸟巢”演唱会人员疏散CA-BP仿真研究  

Simulation research on the evacuation of concerts in Bird Nest based on CA-BP

作　　者：于海明[1] 聂文[1,2] 王云[3] 张磊[1] 白若男[1] 张琦[1] 

机构地区：[1]山东科技大学矿业与安全工程学院,山东青岛266590 [2]山东科技大学矿山灾害预防控制省部共建国家重点实验室培育基地,山东青岛266590 [3]海湾安全技术有限公司,北京100089

出　　处：《安全与环境学报》2015年第6期175-181,共7页Journal of Safety and Environment

摘　　要：针对现今大型活动中的人群踩踏事故频发问题,通过对"鸟巢"演唱会进行人流疏散仿真分析,研究紧急情况下人流疏散逃生规律及均衡疏散方法。通过设置人员速度、移动余值、出口逃生条件等因素扩展元胞自动机行人流模型,并借助Matlab语言实现可视化仿真模拟,以探究人流疏散逃生规律;引入疏散系数,并通过建立BP神经网络获取较优疏散系数,均衡各个出口所承担的疏散任务。结果表明,"鸟巢"演唱会在疏散过程中人群对通行出口的选择较为不合理,仿真疏散总时间长达455.5 s,与理想疏散曲线所得疏散总时间270.0s相比,存在40.7%的偏差,为减小此偏差引入出口疏散系数指导人群定向选择出口,通过仿真模拟得到的疏散总时间为312.0 s,减少约31.5%,出口选择不均衡现象被显著缓解。最后提出了针对疏散不平衡问题的相关措施。The present paper is devoted to a piece of evacuation simulation by taking the Bird Nest in Beijing as a case study sample based on the CA-BP theory.In answering to the urgent need to make successful evacuation in front of the crowd stampede events,we have chosen the topic for exploring such emergency evacuations.In order to reconstruct what has happened in a real life situation in-situ,this paper would like first of all to introduce some indicative parameters concerned,such as human moving speed,human moving residual value,the escaping exits for the urgent evacuation,etc.,in hoping to improve the cellular automata pedestrian model.In addition,we have also managed to accomplish our simulation study by visualizing such a kind of simulated evaluation by using the language of Matlab.Besides,we have also put forward the concept of evacuation coefficient,which helps to achieve a balance between the evacuation task of the exits by setting up the BP neural network.The results of our simulation show that it would be groundless to ask the evacuators to shoose better exits through evacuation simulations.In fact,it wouldn＇ t be possible,nor it be realistic for the evacuators to make a proper choice from the 12 exits given,for there exists a lot of different features from building to building in their architectial features whereas the people to be evacuated tend in most cases to fight for the nearest exits,such as those nearby the areas of A,G,H,M,while some others would be eager to gather near the areas of B,C,E,F,G,L in a short instant,which implies that the safety facilities in the building of Bird Nest fail to be made full use of during the total available time for evacuation,say,455.5 s.In comparison with the result of 270.0 s from the ideal evacuation course,for mere exists a deviation or gap of 40.7%.To reduce the deviation or to make up for the gap,we have proposed the evacuation coefficient to guide the people to choose the nearest exits by using the group of α_n from the 12 exits so as for the evacuators to make

关 键 词：公共安全 安全疏散 元胞自动机 模拟仿真 均衡疏散 BP神经网络 

分 类 号：X923[环境科学与工程—安全科学]