Study On Passenger Flow Warning And Dynamic Performance Analysis Of Subway Station Based On Fluid Queuing Model

With the large-scale planning and construction of urban subway in China and the highintensity increase of its travel demand,the contradiction between the original design capability of subway station and traffic demand is prominent.Accurately grasping the service performance and passenger flow status of the facilities in the station,evaluating the operation risk of the station and determining the passenger flow warning level have important practical significance for formulating a reasonable and effective passenger flow control scheme to ensure the safety of urban subway station operation.The existing researchs on the facilities performance analysis mostly ignore the randomness of passenger arrival rule and facility service time rule.Moreover,the research on passenger flow warning lacks the detailed classification of the real-time status of station passenger flow,and the model of passenger flow warning is more complicated,which can not meet the real needs of grasping the status of passenger aggregation in the station in real time,and the application effect is limited.Therefore,considering that phase-type(PH)can not only infinitely approach any non-negative random variables in theory,but also retain the characteristics of exponential distribution which is easy to handle,the paper introduces PH distribution to describe the randomness of passenger arrival rule and facility service time rule.The fluid approximation technology has high accuracy and timeliness,the paper introduces it to realize the real-time analysis of passenger aggregation in the station.In this paper,the system of passenger service facilities in the station is taken as the research object,from the perspective of system,considering the actual traffic characteristics such as the time-varying demand,system randomness,state-dependent of service ability.The PH-type fluid queuing model at each node of the station,the PH-type fluid queuing network model at the station,the model of simulation,the passenger flow warning research and the facility dynamic performance analysis are carried out.The specific research work of this paper includes:(1)Establishment and solution of PH-type fluid queuing model of the station system.Based on the idea of fluid approximation,the PH-type fluid queuing model of each node facility in the station and the PH-type fluid queuing network model of the station system are established to quantify the complex correlation between the system performance and network parameters.Among them,the PH distribution is used to describe the randomness of passenger flow demand and service capability,the continuous function of time is used to quantify the time-varying passenger demand,and the state correlation of service ability is quantified by embedding macro traffic flow model.At the same time,the existing pointwise stationary fluid flow approximation and numerical integration algorithm are extended to solve the model.(2)Model verification based on discrete event simulation modeling.Taking the Chengdu People’s Park subway station as the case study,by field surveys to obtained geometric parameters of station facilities,and then using the Simulink simulation tool in MATLAB,the Simulink model of each node facility and station network are constructed to verify the accuracy and effectiveness of the fluid queuing model.(3)Research on passenger flow warning and dynamic performance analysis of station system based on dynamic performance indicators.The indexs of travel speed in time range are calculated by the average evacuation time,and the passenger flow warning threshold and warning level of the corridors are comprehensively determined;The indexs of per capita occupied space in time range are calculated by the average occupation space,and the passenger flow warning threshold and warning level of the waiting area are comprehensively determined,so as to realize the timely passenger flow warning of the station system.At the same time,combined with the two indicators of output rate and average queue length,the queue accumulation and dissipation process of the station service facilities are analyzed to realize the real-time dynamic performance analysis of the station service facilities.