Research On The Calculation Of The Maximum Assembling Passengers In Railway Stations Considering Different Arriving Rules

With the continuous development of railway passenger transportation,more and more attention has been paid to the design of railway passenger stations.Among them,the main basis for determining the scale of the station building is the maximum assembling passengers at the station.At present,the calculation method of the maximum assembling passengers is still based on the assembling-coefficient method.In fact,with the current continuous construction and development of high-speed railways,passenger travel needs and travel behaviors continue to change.Whether the traditional assembling-coefficient method is feasible remains to be verified,and whether the value of the empirical coefficient is still applicable remains to be demonstrated.Therefore,it is urgent to propose a more effective and feasible method for calculating the maximum assembling passengers in combination with the characteristics of the current actual passenger flow,in order to provide a more reasonable basis for the scale design of the railway station.Based on the summary of the relevant theories about the maximum assembling passengers at the station at home and abroad,this paper puts forward the calculation method of the maximum assembling passengers considering different arriving rules.According to the actual data,the decision tree is established to predict the arriving rule of passengers corresponding to each train.The passenger gathering curve of the station is obtained by superimposing the passenger aggregation process of each train,and the maximum assembling passengers is calculated.Firstly,the paper analyzes the existing problems in the definition of the maximum assembling passengers.The advantages,disadvantages and adaptability of the existing calculation methods are summarized.Considering that the probability method is scientific and practical,this paper focuses on analyzing and showing the calculation steps of the probability method,including the passenger gathering and dispersing process of single train and the passenger gathering process of station.Based on the purpose of this paper,the shortcomings of the existing probability method are further analyzed and the improvement measures are put forward.Secondly,the calculation database is established,and the arriving rules of passengers of different trains are fitted according to the actual data of existing stations.After comparing the fitting effects of the lognormal distribution,the Weibull distribution and the compound negative exponential distribution,it is found that the lognormal distribution has the best fitting effect and can describe the arriving rules of passengers under different circumstances.Based on this,taking the lognormal distribution as the research object,the values of fitting parameters under different conditions are analyzed.The results show that the parameter values are different when the train departure time,departure and passing type,train number type,the type of city where the station is located,geographical location and subway connection condition are different.In order to effectively predict the arriving rules of passengers for different trains,decision tree model is adopted in this paper.The existing actual train data is divided into training set and testing set.The training set is used to generate the CART regression tree(Classification and Regression Tree),and the testing set is used to verify the effect of regression.The test results show that the regression tree constructed in this paper can effectively predict the arriving rules of passengers.Finally,based on the result of regression tree,the calculation model of the maximum assembling passengers considering the arriving rules is established by using the idea of the probability method.Taking Guangzhou South Railway Station as an example,the maximum assembling passengers in the station is calculated by using the model presented in this paper and the traditional assembling-coefficient method.The applicability and effectiveness of the model studied in this paper are verified by comparing with the actual operating state of the station.