Modeling Train-induced Pedestrian Flows in Suburban Railway Stations
Zoi Christoforou, Ecole Des Ponts ParisTech / LVMT, Bachar Kabalan, , Pierre Argoul,
We develop a microscopic model of pedestrian flows in railway stations in order to estimate subsequent train Dwell Times and to control travel time variability.
Demand for public transport is rising in and around major city centers leading to on-platform and in-vehicle congestion as well as to important delays in scheduled railway operations. A major source of delay is the variability of train travel times and, in particular, the variability of train dwelling at stations. Dwell time (DT) includes the time necessary for doors to open/close and the time needed for passenger alighting/boarding; the latter being mostly subject to uncertainties. Stochasticity is due to a great number of factors including passenger volumes and arrival rate distributions, train headways and operational constraints, the positioning of station-exits along the platform, the positioning of information monitors, and so on. In essence, passenger trajectories interact with each other and with the built environment while being subjected to a specific timing imposed by train operations and acting under limited information.
The objective of this research is to develop a microscopic model of pedestrian flows in railway stations in order to estimate subsequent train Dwell Times. We first estimate passenger demand through a transit assignment model while using scheduled operations as a baseline reference. Next, we model station and train geometry. A specific Origin-Destination pair is assigned to each passenger for his door-to-door journey as well as for the trip leg within the train station (i.e. from station entrance to train door or vice versa). We model passenger flows combining all of the aforementioned data and using a 2D discrete pedestrian model based on non-smooth mechanics and on the Discrete Element Method. Finally, we estimate a new Dwell Time corresponding to the specific boarding and alighting passenger flows. In this paper, we present this model and a French case study that was used for model validation.
The site chosen for model validation is the ‘Noisy-Champs’ RER-A suburban railway station situated in the Paris metropolitan region, France. With over 300 million registered trips in 2011, line A is the European railway line with the highest passenger volumes. Noisy-Champs station is a middle-size station, with a yearly average passenger volume of around 4.7 million. However, in the next years, it will become a major hub serving three new subway lines that are currently being developed. We thus expect to encounter many difficulties in terms of passenger flow management. Furthermore, the RER-A line suffers from serious operational problems. During peak-hours, only 27 out of 30 trains pass through the central part of the line mainly due to Dwell Time variability.
In this paper, we present the results of a series of simulations of a 6-minute long pedestrian flow entering the station, changing level towards the platform, and embarking a train upon its arrival. We statistically compare results to observed Dwell Times for the same passenger levels at Noisy-Champs station. Simulated and observed data are very close allowing us to consider that our model gives realistic outcomes. The developed model can be used for robust train operation scheduling that accounts for variable passenger volumes and headways. It could also be useful to station managers for passenger flow optimization and optimal station design. Finally, it could be used as a decision-aid tool for choosing rolling stock that best fits the stations’ configuration.
Association for European Transport