The numerical CFD simulations are carried out in the tunnel model of 12 m of length, 0.25 m of width and 0.25 m of height. The tunnel model is specified as a 1/20 scale of full scale tunnel. The fire is simulated by a rectangular heat source having a section area of 8 cm × 10 cm. This source is placed at the center of the domain and in the middle of the two sidewalls, with its top surface being set at the floor level. The heat release rate of the fire is changed into different values extending from 1.0 to 6.0 kW. The fire heat release rate is specified as a heat release rate per unit area applied to the source surface and is treated as an average value. The tunnel model is made of ''concrete". The physical properties of this material are specified in the FDS model by the "MATL" command. The tunnel walls are considered smooth. The default velocity condition at the wall surface provided by FDS is assumed. The walls are heated by the radiative and convective heat transfer from the surrounding gas. The two surfaces of the tunnel extremities are both opened to the external ambient environment, but one (tunnel entry) is specified as an air flow inlet in the situation of longitudinal ventilation. The longitudinal ventilation velocity is set up by a supply air at ambient temperature introduced in the tunnel entry surface. The ambient temperature away from fire is approximately about 20 °C in the series of tests simulations.

In order to analyze the effect of vehicular blockage on the backlayering flow spread, a vehicle model representing an obstruction such as a bus or train is placed symmetrically inside the tunnel upstream of the heat source. The vehicle model is a 0.6 m of length, 0.13 m of width, and 0.15 m of height and occupied about 31% of the tunnel section. The obstacle size is selected on the basis of the dimensions values of vehicles at a scale of 1/20. The distance from the vehicle and the source is made equal to 0.1 m. The distance between the bottoms of the vehicle model to the tunnel floor is varied in this study and set at 0, 0.03 and 0.06 m. A schematic representation of considered geometry is shown in Fig. 1.

Series of measuring stations placed at 1 cm below the ceiling of tunnel are fixed in the FDS model to monitor the temperature variation as well as to determine the backlayering length. In effect, the gas temperature distribution below the tunnel ceiling has a sharp decrease at the end of the smoke reverse flow. The presence of backlayering can be easily detected and its length can be measured. The positions of the measuring points of temperature are shown in Fig. 2.