Study On The Network Characteristics Of Urban Rail Transit Based On Intelligent Card Data

Considering the advantage of high transportation capacity,high efficiency,travel convenience,alleviate congestion and environmental friendly,the urban rail transit has been fully supported by the government and gradually become the backbone of urban transportation in recent years.The network has grown continuously in scale and becomes more and more complicated,which makes the urban rail transit step into network operation era.The urban rail transit network is a complex integrity,which is formed by the mutual interaction and influence between the network structure and passenger flow.The planning and operation investment of new lines should abide by the passenger flow requirement,and the network structure should serve only to the purpose of passenger flow transportation.Thus it is one-sided to discuss the urban rail transit network characteristics without considering the transportation feature.Many studies have shown that the complex theory is beneficial for analyzing the logic relationship among each event within the complex system and can be applied on the study of urban rail transit network characteristics.However,most of them only considered pure topological features or the dynamic evolution of the system.Although a few studies have introduced the passenger flow to endow the network more meaning but only some of the static network characteristic values was considered.In this paper,the research work are listed as follows and related achievements were achieved:(1)Surveys on the walking time duration to all of the transfer station with both transfer directions during early peak hours in Beijing urban rail transit network were conducted.which provides the basic parameters for establishing the passenger flow distribution model based on the transfer virtual space.By taking questionnaires in the subway station,the travel preference,the preferred transportation mode after sudden operation accidents,the psychological thresholds(refers to the extra bearable time consuming duration comparing with the original route and increased transfer times due to the accident for the passengers)were obtained,which provides the data support for parameter calibration for the passenger flow simulation model.(2)According to the urban rail transit network characteristics,the L,P space model was established,the statistical characteristic values for the un-weighted and weighted network were defined,and the reliability measure index was obtained based on L space method.For the weights calculation algorithm in the weighted network,the point-line,line-line double layer network model was built as the bottom model,then the shortest path model for urban rail transit based on the virtual transfer space and penalty coefficient according to Dijkstra algorithm,thus the point-to-point path chain structure has been obtained.The edge weight and its distribution were then obtained by assigning the passenger flow according to the transportation intelligent card data.(3)Taking urban rail transit in Beijing for example,the L,P space model were established separately.The network characteristics values(including the node degree or strength,clustering coefficient,shortest path length,node betweenness,edge betweenness,node closeness,adjacent node degree or strength)were calculated for both the un-weighted and weighted network.By assuming several random attack and intentionally attack scenarios the network is cascade destructed and the reliability of the network was analyzed.The reliability index consists of the average path length,global efficiency and maximum connected subgraph.By comparing the differences between the statistical characteristics value and reliability from the un-weighted and weighted network,the influence of passenger flow on the network characteristics is determined.(4)The propagation mechanism of passenger flow in the urban rail transit network was analyzed.The network passenger flow propagation model based on the transportation intelligent card data was established.The model parameter was calibrated by deeply digging the investigation data and the intelligent card data.By combining the passenger flow simulation results with the network structure and line capacity,the identification index for key stations was proposed based on the complex theory.Taking the urban rail transit network in Beijing for example,the procedure for the propagation and dissipation of passenger congestion during early peak hours were simulated;the key stations in the whole network were identified and the distribution law and reason were analyzed.Meanwhile,the overall network reliability under normal operation status at early peak hours were analyzed and compared with the results in the whole day.(5)The sudden operation accident information of Beijing urban rail transit during 2014 and 2016 was counted statistically and the case study scenario for this study was setup.The passengers who were influenced by the accidents were classified according to the questionnaire results,and the choice for each classification of passengers was analyzed.The passenger flow in the whole network after the accident was simulated based on transportation intelligent card data.By comparing the results against the normal operation status,the space distribution of accident influencing classification of each section in the whole network is studied,and the propagation and dissipation law of passenger congestion after an accident were analyzed.Meanwhile,the network reliability under accident occasion at early peak hour was analyzed and compared with the results under normal operation status at early peak hours.