Numerical Simulation of Fire Growth, Transition to Flashover, and Post-Flashover Dynamics in the Dalmarnock Fire Test

Abstract

The objective of the present study is to evaluate the ability of current Computational Fluid Dynamics (CFD) tools to simulate compartment fires with flashover followed by under-ventilated and/or quasistoichiometric, partially-under-ventilated conditions. Current CFD capabilities are illustrated using the Fire Dynamics Simulator (FDS, Version 5), developed by the National Institute of Standards and Technology, USA. The FDS modeling capability is evaluated by detailed comparisons with an experimental database previously developed by the University of Edinburgh, UK. The test configuration corresponds to a fullscale fire test known as the Dalmarnock fire test (test 1). The description of the flammable content in the fire room is based on a standard modeling approach in which the ignition time of flammable objects and materials is calculated using a local heat transfer solver, while the fuel mass loss rate after ignition is prescribed using experimental data from cone/furniture calorimeter tests. The simulated Dalmarnock fire scenario includes flashover, a first post-flashover stage that is under-ventilated and characterized by burning outside the fire room, and a second post-flashover stage that is partially-under-ventilated and characterized by distributed burning inside the fire room. Transition to this second stage is triggered by window breakage in the fire room. The different stages of the fire scenario are analyzed in terms of the fire room global equivalence ratio (GER), which is considered as the main controlling parameter of the fire behavior. Comparisons between numerical results and experimental data are relatively good when considering the global features of the fire dynamics, e.g., the time history of the spatially-averaged heat release rate. Comparisons are not as good when considering local features, e.g., the time history of gas or wall temperatures, or that of wall heat fluxes.