Abstract: When a subway train is forced to stop in an interval tunnel due to an emergency, passengers inside the carriages may need to evacuate in place using the side platforms. Considering the influence of fire source on passenger movement behavior, this paper introduces a static field and a fire source field to describe the attractive effect of exits and the repulsive effect of the fire source on passengers, respectively, and develops a cellular automaton model for passenger evacuation in subway tunnel train fire scenarios. Based on this model, the effects of factors such as fire source power, fire source location, door opening/closing status, and evacuation direction on the evacuation efficiency using side platforms are investigated. The results show that, compared with non-fire scenarios, evacuation time in subway train fire scenarios is generally prolonged, and the difference between the two increases with the number of evacuees. As the parameter Kf, which characterizes fire source intensity, increases, the evacuation time first decreases slightly and then increases continuously, with the growth rate increasing as Kf rises. When passengers evacuate unidirectionally toward the front of the train, the evacuation time gradually increases as the fire source location moves from the rear to the middle of the train. After closing the two doors closest to the fire source, the evacuation time in the train-middle fire scenario is reduced by approximately 40%. Furthermore, when a fire occurs in the middle of the train, the evacuation time under bidirectional evacuation is only one-third of that under unidirectional evacuation toward the front in a train-head fire scenario.

Key words: metro interstation tunnel evacuation, cellular automata, evacuation simulation, lateral platform evacuation