Emergency Evacuation and Safety in Complex Environments

Effective egress system design and management are both fundamental in order to guarantee people's safety in case of compartment fires. In this thesis PASS (Preliminary Assessment of the egress System Safety), a method that allows a rapid screening of the egress system vulnerability, is proposed as a complementary tool to traditional safety assessment methods. PASS is a set of simple equations that allows assessing at different levels of analysis the egress system performance without simulating the evacuation process. Its equations incorporate factors for including in the analysis the main interrelationships between people-egress system-environment, and for linking the two main concepts of fire sciences: ASET and RSET. In PASS, ASET is estimated through the use of standard times, and simple tools were developed in order to allow their rapid calculation. The results obtained with these tools were compared with results from the computational fluid dynamics model FDS, and a good agreement was found for a set of reference scenarios. Some factors adopted in PASS to consider people-egress system interrelationships were quantified through experiments carried out in the novel laboratory LabCUBEegress. LabCUBEegress is a movable laboratory with a flexible configuration which can be customised to create several layouts and to reproduce different parts of an egress system. It was conceived and designed in order to quantify human behavioural factors. A set of experiments involving a wide cross-section of population were performed and the effects of exit signs on people's choice at a T-intersection were investigated. The results revealed that with a left exit sign only 70% of individuals choose the shown direction and this value drops to 60% for groups. These results were used to calibrate a PASS factor and were eventually integrated into the method. Finally, PASS was applied to a case study. Its results were compared with two evacuation models (STEPS and FDS+Evac) and data collected from an announced evacuation drill. A good agreement between PASS analysis, experimental and numerical data was obtained for the scenario studied (a secondary school). Thanks to the results achieved, PASS can be proposed as a simple and rapid stand-alone tool for the pre-design and management of the egress system safety. For extending the range of validity of PASS, in future works the method may be tested in different scenarios.