Modelling factors that influence CFD fire simulations of large tunnel fires

A series of CFD fire predictions of a 100 MW Memorial tunnel fire test are performed to investigate several factors that influence the outcome of CFD fire simulations of large scale tunnel fire scenarios. This study shows that at the critical ventilation velocity, the occurrence of back-layering is very sensitive to factors such as variation in ventilation velocity, tunnel surface condition, the width of the modelled fire source and the presence of relatively minor obstacles in the vicinity of the fire source. A serious question has also been raised over the suitability of the use of inlet boundary conditions to represent the ventilation flow generated by jet fans. To a large extent, the success of simulating a tunnel fire test such as the Memorial Tunnel test relies on correct replication of the test conditions and setup. However, for most large scale tunnel tests, not all the information required to correctly specify a CFD fire simulation is available and hence sensitivity analyses over critical model setup parameters is essential to correctly interpret differences between predictions and experimental measurements