Study On The Forecast Of Urban Rail Transit Passenger Flow Considering Severe Weather

Passenger flow is an important basis for new line design,infrastructure construction,vehicle purchase,personnel allocation and operation management of urban rail transit.In severe weather,the station passenger flow fluctuates greatly,which brings great hidden danger to the operation organization and passenger safety of urban rail transit.The study of urban rail transit passenger flow in severe weather and the establishment of passenger flow prediction model can provide certain data support for the formulation of traffic emergency plan in severe weather.In this paper,based on the BIGEMAP platform,using ordinal and virtual variables of meteorological data processing methods,through the way of stepwise regression,the relationship between the size of passenger flow and various factors under the influence of bad weather is analyzed,and on this basis,the passenger flow prediction model is constructed:1)According to the 0.5h granularity time-sharing passenger flow data of Beijing urban rail transit station,the distance evaluation function is constructed to determine the optimal clustering number.On the premise of meeting the minimum distance evaluation function value,the minimum similarity and the maximum difference between categories,the station is divided into six categories,namely,commuter residential type,commuter partial residential type,commuter balanced type,commuter partial commercial type,all day balanced type,commuting business type,and the time-sharing passenger flow characteristics of different types of stations are significantly different.2)Based on the nine item average method to process the passenger flow data,the smooth value of the passenger flow in bad weather is obtained,and the time-sharing passenger flow curve and the passenger flow volatility are compared and analyzed.The research shows that the daily passenger flow of Beijing urban rail transit network decreases by about 400000 due to the rainfall weather;the difference of the characteristics of various stations results in the different changes of the passenger flow in rainy weather,and the peak passenger flow increases in the morning and evening at the first and sixth subway stations In addition,the passenger flow of category 5 stations in each period is significantly reduced;the impact of air pollution on the passenger flow of road network and stations is relatively small.3)By analyzing the characteristics of passenger flow changing with time and space,the influencing factors of passenger flow in the three dimensions of severe weather,the nature of land around the station and the clustering level of the station are determined.Based on the topological calculation of meteorological cloud map and map,the coverage of severe weather is determined.Based on the data of different land areas obtained by BIGEMAP platform and the clustering analysis of the station,the sequential variable processing of rainfall weather is carried out,on the basis of the virtual variable processing of haze weather,the stepwise regression method is used to build the rail transit passenger flow prediction model considering the bad weather.4)Taking Beijing Metro Line 4 as an example,six groups of urban rail transit passenger flow prediction models considering the bad weather and different passenger flow characteristics in different periods of time in Beijing are established.By comparing the passenger flow predicted by the model with the actual passenger flow of the station,the model has a good fitting effect.