The simulation of crowds in emergency situations, e.g. evacuation of passenger ships, has recently gained an increasing interest. Such simulation are used to scrutinise the access of rescue zones and optimise the design of escape routes in order to ensure a fair survivability-rate and a successful evacuation process in case of severe accidents.
Accordingly, different modelling strategies exist to analyse the behaviour of individuals (aka. agents) in emergeny situations. The present research is based upon detailed microcopic models for the behaviour of each individual and utilises social forces to determine the agent movement. Special emphasis is given to the introduction of endurance/exhaust aspects to the agents' capabilities. Moreover, the interaction with fluid dynamic processes, i.e. the evolution of stack gas or the water ingress, is modelled.
The methodology refers to a particle-based simulation technique (SPH-Code GadgetH2O). The particles mimic the behaviour of different continua (water, gas) as well as human beings (agents) and their interaction. Due to the 3D transient nature of the approach a detailed analysis into the specific design of escape-route elements can be performed.
Fig. 1: Crowd simulation for a group of agents separated by one room and two doors from the rescue zone (left to right: inital, 50s, 90s, 130s, 180s, 270s, 370s)
Internal project funded by Hamburg University of Technology (TUHH)