Research On Calculation Method Of Rail Car Congestion Degree Considering Station Type

With the continuous expansion of the urban rail transit network and the frequent occurrence of urban road traffic congestion problems,rail transit,as one of the methods of public transportation with greater advantages in terms of punctuality and convenience,has been favored by many urban residents.More and more urban residents are willing to choose rail transit for travel.However,with the continuous increase in travel demand,the congestion of passengers in the rail network section is becoming more and more serious,especially when the passenger flow is concentrated in rush hours,and the rail network section is crowded.More prominently,it is of great significance to provide the actual service level of urban rail transit by studying the congestion of urban rail sections in the future.This thesis mainly processes and mines the AFC ticket data,establishes an LSTMbased urban rail network interval traffic short-term prediction model,combines the land attributes and the characteristics of the passenger flow in and out of the station to classify the station types,and proposes a classification based on the station type The calculation method of the congestion degree of the carriages in any interval section,and finally the suggestion of adding the interval car to a single line of urban rail is given.The main research contents of this thesis are as follows:(1)Analyze the AFC ticket and card data,and explain the characteristics and influencing factors of compartment congestion.The root cause of compartment congestion is the mismatch between rail passenger flow and compartment capacity.The main reasons include passenger types,etc.;(2)Constructed an LSTM-based short-term prediction model of urban rail transit traffic.Considering from the perspective of the network,taking the passenger flow data of each station of the rail network as input,the short-term flow forecast of 184 sections taking the Chongqing rail network as an example is carried out,and four predictions are drawn from the long-term sequence The task is tested,and the results show that for the passenger flow data of orbit in and out of the station,the short-term prediction task with time window 2 and 3 proves that its correlation is stronger,and the model with time window 3 is better for forecasting during peak periods.The model with a time window of 2 is better for forecasting during peak periods;(3)A calculation method for compartment congestion based on station types is proposed.Combining the map for land planning and the characteristics of the passenger flow in and out of the station,the rail stations are divided into five types,and the passengers of the five types of stations are classified through on-site sampling surveys,and the proportions of the eight types of passengers and the corresponding vertical projection area are calculated.Then obtain the per capita seat area based on the station type;then the program automatically calculates the number and types of stations in the interval radiation range,associates the station passenger type with the interval passenger type,and divides the interior of the compartment into five types of areas according to the standing preference of the passengers,and the comprehensive weighting calculation is different The effective use of the area of the types of trains;finally,the method for calculating the congestion degree of compartments based on the station type is proposed in this thesis,and the traffic of each section of the output network of the prediction model is substituted,and finally the congestion level of the sections of the Chongqing rail network is evaluated.The results show that the areas around residential and work stations are less crowded,while the areas around shopping and sightseeing stations and external hub stations are more crowded.