The Sensitivity Analysis On Passenger Flow Of The Urban Rail Transfer Station

Prediction of urban rail transit passenger volume is the base of rail transit construction and design, passenger flow forecast result directly affects the traffic infrastructure and auxiliary facilities configuration. But the result of forecast and the actual data often exist great bias due to many uncertain factors, such as land use, population, transportation policy fare and so on. The transfer station as the supporting point of the rail transit network skeleton structure, which passenger flow structure is more complex, the accuracy and reliability of forecast result is more difficult to grasp. Sensitivity analysis can find out sensitive factors and its influence on passenger flow, for the early stage of the late in the transfer station design planning and operation management to provide the reference basis.This paper analyzes the uncertainly source of rail transit passenger flow, determines the influencing factors and quantify the parameters, on the basis, classifies the affecting factors from the qualitative and quantitative, internal and external aspects, using fuzzy comprehensive evaluation method and Pearson chi square to extract the sensitivity factors respectively on the qualitative and quantitative influencing factors. To determine the sensitive factors as:fare system, system service level (interval), population scale (distance of transfer), network structure, and bus interface.The sensitivity analysis is based on the analysis of the application of BP neural network. A sensitivity analysis model of transfer station passenger flow is proposed and constructed on BP neural network. Sensitivity factors and transfer site as the initial input variables, and then using Matlab neural network design, build the sensitivity analysis model of the station transmits flow and transfer passenger flow.Transfer station passenger flow prediction of Chengdu Metro Line7as an example, built and verified the model, get the transfer passenger flow and the station transmits flow base on sensitivity analysis.