Evacuation Simulation Modelling For A Deep Underground Subway Station

Pathfinder, amongst other agent-based evacuation modelling software, contains an underlying approach that all agents are fully aware of their location with respect to all available exits in the evacuation model. This may or may not be true in real life. Building occupants may not be fully aware of all available exits in a building, particularly when exits are placed within less frequented areas (e.g. back-of-house facilities in a hotel or shopping mall). There is also an inherent tendency for building occupants to evacuate a building via routes/exits that they are familiar with. Agent-based evacuation simulation models have proved to be both useful and representative of occupant evacuation for typical multi-storey high-rise buildings. NFPA 130 requires that all occupants evacuate from the station platform within 4 minutes, with an underlying assumption that all exits will be utilized equally. In our case, evacuation modelling for a deep underground subway station showed mixed results. User judgment is then critical in assessing the sensibility of the evacuation simulation results and various amendments to the inputs would be required to arrive at a more logical result. This paper examines an evacuation simulation study carried out for a deep underground subway station, comparing differences in various user input parameters (e.g. various cost factors for occupant profiles, locations of final exits etc.) across the same simulation model, finally arriving at what we judged as being sensible. The key challenge faced was modelling the locations of the various final exits, simply modelling the actual layout of the station with the actual locations of the final exits gave somewhat skewed results. Judgment in making sensible changes to the model is required to ensure results obtained made adequate sense.