He-in Cheong - Imperial College London
When assessing evacuation from a building fire, computational fluid dynamics (CFD) models of fire can aid with the calculation of available safe egress time (ASET) and evacuation and crowd dynamic simulation can help estimate the required safe egress time (RSET). Some existing models and simulation environments directly couple the fire and evacuation models for a holistic assessment. These, however, often use the measure of incapacitation in terms of fractional effective dosage (FED) and/or fractional irritant concentration (FIC) to quantify tenability condition and to affect the walking speed of the agents. Much fewer models use the measure of visibility to predict the walking speed and exit/route selection of the agents. Even when visibility is taken into consideration, often the local level at the position of the agents are used, especially when the exit door is not within the line of sight. This paper presents a new and computationally resourceful method of analysing visibility along the entire line of sight for each agent through the consideration of soot level. The method is tested on the agent-based crowd simulation tool, Oasys MassMotion version 10, using its software development kit (SDK). A simple geometry of 50 m by 50 m floor with walls is used to test evacuation in three ways: 1) without any coupling with the fire model, 2) using local visibility values, and 3) through the consideration of soot level analysis along the line of sight for each agent. The results are compared to show how the new method can provide a more realistic coupling of CFD and evacuation models and can affect the overall RSET.