THE USE AND PREDICTIVE CAPABILITY OF CFD FOR FULLY DEVELOPED NATURAL FIRES IN LARGE COMPARTMENTS

Abstract. The purpose of the present study is to explore and validate the efficiency and accuracy of the predictive capability of CFD models for fully developed natural fires in large compartments. The FDS CFD software was used to simulate three large-scale fire tests. The fire tests (nine in total) had been carried out in 1993 at Cardington UK, in collaboration with BRE/FRS and CORUS (British Steel at that time). The main parameters of the three chosen large-scale fire tests are the compartment's dimensions, the fire load, the ventilation opening, and the ignition type. The duration of the fire tests, starting from the ignition until the extinction of the fire, was between 70-120 min and the fire source was wooden cribs. The simulations were in full time and full scale. The FDS model solves numerically the well known Navier-Stokes equations with emphasis on smoke and heat transport from fires. It uses a mixture fraction combustion model assuming that combustion is mixing controlled, and that the reaction of fuel and oxygen is infinitely fast. Turbulence is treated by means of the Smagorinsky form of Large Eddy Simulation, while thermal radiation is computed using a finite volume technique. Despite the assumptions made and the complex phenomena involved, the correlation between the computational and experimental results was in very good agreement. The fire spread was modelled accurately and the difference for the peak temperatures was 5%-20%. However the results are sensitive to some input parameters and so each fire scenario should be treated individually for a robust and accurate solution