Abstract
Evacuation path of pedestrians at subway station can directly affect the evacuation efficiency, and then affect the service level of the station. A Fisk stochastic user equilibrium model considering the congestion factor is established to assist evacuation path planning, which could avoid the bottleneck area and dangerous situation in advance. By comparing the field statistical data with simulation data of the number of pedestrians at the entrance gate of subway station, it is verified that the social force model and the minimum cost model can basically reproduce the movement law and path selection behavior of pedestrians at the subway station. Simulation experiment is carried out to compare the evacuation efficiency at the subway station under the stochastic user equilibrium model and the minimum cost model. The results indicate that the number of pedestrians at each gate is relatively balanced which can avoid overcrowding under the stochastic user equilibrium model. The total evacuation time can be reduced by 15%, and the individual evacuation time can be saved about 0.57 min. By comparing the 3D overall pedestrian trajectory and the 2D local pedestrian trajectory, it is found that the stochastic user equilibrium model is of great significance for evacuation path optimization at the subway station. The quantitative relationship between the number of pedestrians and evacuation time at the subway station is given through regression analysis under stochastic user equilibrium model and minimum cost model. The suspension of escalators can obviously result in congestion at the gate and increase the total evacuation time. When pedestrian density at the station is low, closing a certain exit can significantly improve the pedestrian traffic efficiency at the gate and greatly reduce the total evacuation time. The stochastic user equilibrium evacuation path planning model constructed in this paper provides a guidance strategy for pedestrian evacuation, which could improve the efficiency and safety of subway evacuation system. (C)& nbsp;2022 Elsevier B.V. All rights reserved.
NSTL
Elsevier