Theory And Method Of Passenger Flow Prediction Based On Urban Rail Transit Station Peak Deviation

The urban rail transit station is the only node of the relationship between the rail system and the outside world.The passenger flow of urban rail transit station is the "source" of the passenger flow of urban rail transit line and line network,which reflects(produces)the corresponding passenger flow of line and line network.At the same time,it also represents the passenger flow in and out of each station.In other words,the superposition of ridership of stations formed line passenger flow and network passenger flow.Thus,the station’s peak hour is not exactly the same as the line’s peak hour.The line’s peak hour can only represent the peak hour of most stations on this line,not the peak hour of all stations,which results in a station peak deviation.However,in terms of current urban rail transit passenger flow forecasting work,in order to control network errors,the line’s peak hour is often selected,and the station’s designed ridership is also the ridership of the station during the line’s peak hour.For stations whose ridership has a big difference between the station’s peak hour and the line’s peak hour,the station design capacity is often insufficient.In addition,the greater the difference of the station’s peak hour,the more difficult it is to select the time parameters of the line’s peak hour.Therefore,the accurate identification of station peak time is of great significance for the design and operation of urban rail transit station itself,and for the time parameter selection in the peak time prediction using the macroscopic passenger flow prediction model.Firstly,through the analysis of domestic and foreign research results,this paper summarized the development process of the station passenger flow prediction method Meanwhile,the research status of station time distribution and peak passenger flow was also analyzed.Then,this paper analyzed the difference between the station’s peak hour and the line’s peak hour in different cities and different rail transit network periods in one city,and illustrated the universality of the inconsistency between the station’s peak hour and the line’s peak hour.Since the ridership during station’s peak hour belongs to the research scope of the time distribution of station,this paper found influencing factors of station peak hour deviation and station peak ridership deviation according to influencing factors of time distribution.On the basis of the time and space relationship in the process of residents’ travel,the formation mechanism of the station’s peak hour was discussed.According to the formation mechanism of the station’s peak hour,the station’s peak hour and ridership prediction model was established.Finally,it took Xi ’an as an example to study,and discussed the impact of office crowds in noncommuting land on the station’s peak hour.The main work of this paper is as follows:(1)This paper discussed the relationship between the line’s peak hour and the station’s peak hour,and found influence factors which can offset station’s peak hour according to the factors that influence factors of time distribution and peak hour of station.The analysis method of characteristic decomposition was used to analyze the time distribution of ridership of the station,and it was found that the basic form of the time distribution of ridership of the station on weekdays was a "saddle shape" with obvious morning and evening peak characteristics for both boarding and alighting ridership.The peak hour of difference form was inconsistent with the peak hour of the basic form.After different proportions of the two distribution forms were superimposed,different ridership distribution patterns and different station peaks would be formed.The proportions of the two distribution forms of each station were related to the surrounding land and station location.According to the results of the factors affecting the time distribution of station ridership,a geographically weighted regression model was used to explore the factors affecting the deviation of station peak hour and ridership during station peak hour in different cities or one city during different line network structures.The results found that the station peak deviation of different cities was related to the land attributes around the station and the location of the station,but the different line network structure had little effect on the station peak deviation.Finally,the extra peak hour factor and the station peak deviation coefficient was discussed.(2)Based on the general rules of residents’ travel process,this paper analyzed the continuity relationship of residents’ travel chain,residents’ estimation of their travel departure and arrival time,and the connecting characteristics of rail transit.According to the difference of residents’ estimation of departure and arrival time between morning and evening rush hours,the formation mechanism of boarding peak,alighting peak and transfer peak of station was discussed in the morning and evening peak hours.During the morning peak hour,the land around the station had a greater impact on the alighting ridership of the station.The time distribution of alighting ridership was the superimposed result of the time distribution of travel destinations attracted by land with different attributes.The boarding ridership and the transfer ridership of the station were the reverse of the alighting ridership in the rail transit network.Therefore,the boarding ridership and the transfer ridership of the station were the weighted overlay of the alighting ridership for different travel purposes.The difficulty of forecast for station’s evening peak ridership was the determination of the origin attribute of the trip for the purpose of "going home".According to the continuity of the locations of th e residents’ travel chain,the land attributes for the purpose of "going home" could be found.In the mechanism research,it was found that the weight of ridership distribution between boarding and alighting stations was related to the degree of correlation between stations.Referring to the spatial autocorrelation index Moran’s I index,the spatial correlation test index for the ridership relationship between stations was constructed,and it was found that the rail stations had spatial correlation,and the correlations are different.Refer to the gravity model,it constructed a calculation method for ridership distribution between boarding and alighting stations.(3)According to the formation mechanism of urban rail transit station peak,the prediction models of station boarding peak,alighting peak and transfer peak were constructed respectively.morning alighting peak time and ridership forecast model was established based on the station direct estimation model,considered the time distribution of different trips for different purposes,and the time distribution function was introduced into the model.Calculated the definite integral of this function,that was,found a certain time,so that the time when the station alighting ridership reached the maximum in a given period,and this time was the starting time of the station peak hour.The result of the definite integral was the station alighting peak ridership.(4)The station boarding peak forecast model was based on the station alighting peak forecast model,and was established by according to the characteristics of the time distribution function caused by residents traveling in rail transit.And introduced the attraction coefficient of ridership between stations in the model.The attraction coefficient of ridership between stations was established according to the gravity model,through the complementarity between the land attributes of boarding and alighting station,the land price difference,and the distance distribution probability of residents traveling for different purposes using urban rail transit.