Road and railways tunnels are constructed for decreasing the transportation time inside city or intercity. The fire safety systems are mounted for the safe use of tunnels. Therefore, it is important to accurately predict the fire-induced air velocity,temperature and smoke concentrations in tunnel fires in order to design efficient fire protection systems. To this end, scaled tunnel models are used and experiments are carried to understand the phenomena on these tunnel models. In addition, the studies for investigating the tunnel fire phenomena and their methods of modeling techniques for fire experiments are mentioned. In the literature, there is no sufficient information about vehicles&#039 / blockage effect on heat release rate and temperature distribution inside tunnel with different ventilation velocities. As a result, in order to research this subject, the scaled model tunnel is constructed in Fluid Mechanics Laboratory. Based on the Froude number scaling, wood sticks with different configuration inside the model tunnel are burned in a controlled environment. The heat release rate measurement, sampling of gases after combustion, mass loss rate of burning models and temperature distribution along the tunnels with different longitudinal ventilation velocities are measured to investigate the effect of different cross-sectional areas of the burning substances. Furthermore, the model vehicles having a square base area are built according to wood crib theory. The results are investigated with statistical techniques called &quot / Analysis of Variance&quot / and general results have been tried to be reached. It is determined that the variation of air velocity inside tunnel is not so effective, but model vehicle&#039 / s cross sectional area is directly proportional to heat release rate.

L’objectif de cette thèse est de présenter des critères simples pour l’évaluation des performances d’une ventilation transversale. On définit l’efficacité et le rendement, calculés à partir du flux de déficit de densité créé par l’incendie. Après une étude préliminaire permettant de caler les modèles numériques, l’influence de différents paramètres sur ces valeurs a été évaluée. Dans un premier temps, l’étude a porté sur une seule trappe et a montré que le courant d’air dans le tunnel est le paramètre le plus important. A l’inverse, la forme de la trappe pour une surface donnée n’a que très peu d’importance, sauf dans des cas limites (fente sur toute la largeur du tunnel). L’étude d’un canton à huit trappes confirme l’importance du contrôle du courant d’air. Lorsqu’on diminue le nombre de trappes en conservant la même surface totale d’aspiration, cette sensibilité tend à disparaître, mais l’efficacité diminue. Enfin, l’étude de l’influence de l’encombrement montre une légère amélioration de l’efficacité, due à l’obstruction d’une partie de la section par des camions, qui bloquent les fumées au droit des trappes. / The aim of this study is to present simple criteria to evaluate the performances of a transverse ventilation system. Efficiency and yield have been defined, based on the flux of density deficit createdby the fire. After a preliminary study of numerical models, the influence of different parameters on those criteria has been investigated. First, the study focused on the simplified case of a single damper. We showed that the air flow speed in the tunnel is the more important parameter. On the contrary, the shape of the damper has little influence on the efficency, except in unusual cases (slit on the whole width of the tunnel). Then, the study of a longer portion of the tunnel confirmed the prominence of the limitation of air flow speed. Comparison with a case with fewer dampers but the same total extraction section showed that this sensibility to air flow speed tends to disappear for larger dampers, but that the efficiency is lowered. Finally the study of the influence of heavy-goods-vehicules obstructing the tunnel reports better results, due to the fact that big objects tend to confine smokes.