Publication Type: Conference Paper
Source: Pedestrian and Evacuation Dynamics PED (2012)
When evacuating a high-rise building, two streams of evacuees meet at the landings: those who are already descending the stairs and those who are about to enter the staircase through the landing door. In this article, merging of these streams is studied with agent-based simulations using both queuing model and cellular automata.
The queuing model has a first in first out (FIFO) queue on each floor for agents who are about to enter the staircase. The landing has sufficiently space for the agents to enter according to the merging priority. The number of agents in the landing is limited by a maximum density and the flow is limited by a maximum flow, which depends on the width of staircase. In addition, the landing door width on each floor limits the flow of entering agents.
The cellular automata models movements of the agents in detail. This model does not use any fixed merging priorities but the joining is a result of simulated interactions between the agents and the building.
An evacuation of a high-rise building is simulated with both the queuing model and the cellular automata model. For the queuing model, the following merging priorities are used:
1) Stream in the staircase has a priority: agents waiting at the floors have to wait until the staircase is empty;
2) Stream from the floors has priority: agents in the staircase have to wait until the floor is empty;
3) Both streams have equal priority.
The merging priority does not have much effect to total building egress time. In the first case, the upper floors are evacuated more quickly than in the other two cases. The merging priority affects to the queuing and waiting times on the floors. Results of the cellular automata model are compared to queuing model with different merging priorities. Similar agent characteristics are used, but in the cellular automata model, the shapes and sizes of landings are varied. The effect of the landing shape to the merging priority is compared.