Real-time Estimation Model And Control Method For Urban Rail Transit Line Load

Under the condition of network operation of urban rail transit,the complexity of the spatial and temporal distribution characteristics of passenger flow and the complexed match relationship between the demand of passenger flow and the transport capacity make higher demands on lines' and stations' passenger flow organization and safety assurance.Accurate and timely calculation of passenger load on all trains and station infrastructures can provide real-time information about passenger flow distribution for the operations management department,and thus achieve passenger flow early warning and emergency response.Based on passenger flow as the main line,and on a basis of using the AFC data of rail transit,the real-time passenger flow prediction model is constructed by real-time estimating the OD of passenger flow and calculating the parameters of pasengers'entering and exiting at the stations as well as passengers' transfer walking processes.The model can calculate real-time parameters such as full-load ratio between sections during peak hours,the pass-ups probability of waiting passangers on the platform that cannot get on the former train,and passenger traffic congestion probability of station infrastructure.The model can also find the bottleneck of congestion in the line and station equipments,and establish a reasonable flow control scheme for entering and transfer-in passenger flows.The detailed research contents are as follows:(1)The estimation of passengers' whereabouts after swiping cardBased on the fact that the stability of entering passenger flow rate in rail transit station is highly correlated with the statistical time scale,this paper establishs multiple-model combined estimation method under different time scales and studies.The influences of different time scale combinations and the types of land used around the station on the estimated results are studied.The case study shows that at morning and evening peak period,compared with the single time scale estimation method,the 15-min time division rate estimation error of the multi-time scale combination model proposed in this paper is reduced by nearly half.This shows that the results of this paper can provide more accurate basic data for real-time station and line passenger flow calculation.(2)The parameter estimation of passengers' passing in and out and transfer walking processConsidering the pulse characteristics of passenger flow arrival in rail transit,a queuing network model of passenger flow in subway station based on discrete time Markov chain is established,and the influence of passenger arrival form and station transfer facility layout on passenger flow prediction is studied.The results of the case show that the calculation error of the model is smaller than the actual survey data.Compared with the existing literature,the calculation error of the passenger flow density at the peak period of this model is decreased.Compared with the existing commercial software,the CPU operation time of this model is significantly reduced under the premise of satisfying the accuracy of a certain passenger flow calculation.(3)Real-time calculation of the load of train and station infrastructureConsidering the successive passenger flow and the pulse transfer passenger flow,the real-time passenger flow calculation model and platform are constructed.Compared with the existing literature which considers the station as a single node simplifying the process of passengers' entry,exit and transfer,the calculating platform in this paper established a complete shift process of passengers including swiping card,walking through the equipment in the station like escalators and elevators,waiting on the platform,getting on the train and getting off for transfer or leaving the station,which can calculate in real time the full load rate among the trains,the pass-ups probability of the passenger waiting at the station,the blocking probability of the passenger flow in the station infrastructure and so on.(4)Off-line control optimization of train and station infrastructure loadBased on the established passenger flow prediction platform,a flow intensity optimization model for passenger flow and transfer station is established according to the distribution characteristics of passenger flow load.This paper studies the influence of the adjustment of the control flow intensity on the inbound and transfer passenger flow on the full load rate of the train and the pass-ups probability of the waiting passengers on the platform at the early peak period.The results of the case show that when the suburban line passenger flows transfer into central urban line at morning peak,taking flow-limiting measures on inbound or transfer passenger flows at 3 sorts of stations-the end point station of the line,the transfer station and the middle station adjacent to the transfer station,can help increase the equalization of the connection trains' capacity usage in central urban area.(5)On-line control of the load of the train and station infrastructureConsidering that the flow control scheme of off-line optimization may not meet the needs of real-time passenger flow,a real-time computation method of flow control scheme based on real-time AFC data is constructed.First,the heuristic algorithm is used to get the control scheme online quickly.Then,the flow control scheme obtained online is adjusted based on the former historical flow control scheme at the same time period.The results of the case show that the effectiveness of the passenger flow control obtained by this method is better than that by using the off-line optimization of the passenger flow control scheme which only uses the historical passenger flow data off-line.At the same time,the online solution time of the calculation method can meet the real-time passenger flow prediction requirements,and the flow control scheme can guide the operation management department to take effective passenger flow control measures in time.