Fire disasters happen every day all over the world. These hazardous events threaten people's lives and force an immediate movement of people wanting to escape from a dangerous area. Evacuation drills are held to encourage people to practise evacuation skills and to ensure they are familiar with the environment. However, these drills cannot accurately represent real emergency situations and, in some cases, people may be injured during practice. Therefore, modelling pedestrian motion and crowd dynamics in evacuation situations has important implications for human safety, building design, and evacuation processes. This thesis focuses on indoor pedestrian evacuation in fire disasters. To understand how humans behave in emergency situations, and to simulate more realistic human behaviour, this thesis studies human behaviour from fire investigation reports, which provide a variety details about the building, fire circumstance, and human behaviour from professional fire investigation teams. A generic agent-based evacuation model is developed based on common human behaviour that indentified in the fire investigation reports studied. A number of human evacuation behaviours are selected and then used to design different types of agents, assigning with various characteristics. In addition, the interactions between various agents and an evacuation timeline are modelled to simulate human behaviour and evacuation phenomena during evacuation. The application developed is validated using three specific real fire cases to evaluate how closely the simulation results reflected reality. The model provides information on the number of casualties, high-risk areas, egress selections, and evacuation time. In addition, changes to the building configuration, number of occupants, and location of fire origin are tested in order to predict potential risk areas, building capacity and evacuation time for different situations. Consequently, the application can be used to inform building designs, evacuation plans, and priority rescue processes.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:626669 |
| Date | January 2014 |
| Creators | Roan, T. |
| Publisher | University College London (University of London) |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://discovery.ucl.ac.uk/1417798/ |
Page generated in 0.0011 seconds