Multiscale approach for the study of ventilation system in Frejus tunnel in case of fire

The complexity in the analysis of fire evolution requires many years before the complete definition of a device able to simulate and predict them behaviour. Since the evolution of numerical simulations, there was the necessity to create a software able to analyse critical conditions generated by fire studying how control it. Many aspects must be considered such as the reaction of the species or the effects of the external environment on the fire itself because they can change its evolution in time. The case study considered in this research is the Frejus tunnel. It is one of the longest transalpine tunnels with a transversal ventilation system due by its length. The history shows many critical accidents connected with fire and they show particular effects difficult to be controlled as in the case of Mont Blanc accident in 1999. Moreover, one of the things that is highlighted studying older accidents is that people in danger are not only who is directly connected with the vehicles enrolled in the accident itself, but many times one of the worst cause of death is connected with intoxication and as a consequence with smoke propagation inside of the tunnel. For these reasons the smoke evolution after fires becomes a very important point in the research because it is possible at least to limit and control the effects generated by an accident trying to confine as much as possible the smoke and as a consequence also the temperature in the tunnel. The main goal of this job is to focus in a deeper way on the analysis of the fire in Frejus tunnel caused by a vehicle inside of it trying to define certain guidelines about how the ventilation system has to react in order to confine the propagation of the smoke. The simulations are performed with the help of a particular method called Multiscale approach. The basic software that is used is Fire Dynamic Simulator (FDS) which is born specifically for the analysis of fire in closed environments. It allows to simulate a 3D space with a fire inside and through analytical equations implemented, it is able to describe the evolution of fire and of the smoke considering all the factors that can affect them such as the wind or the combustion processes. The software is based on the common Navier-Stokes-Fourier (NSF) set of equations with the addition of other physical laws to also consider the effects of combustion. The main problem of FDS is connected with the real time of the simulation that is required to get the results. In fact, the higher is the number of cells the longer will be the real time to obtain the solution because the software has to perform a higher number of calculations. This is why multiscale approach is introduced in simulations of this type because it can avoid problems connected with time. The time problem becomes more relevant when the tunnel considered has a length almost equal to 13 km which for sure requires an incredible high number of cells to be simulated entirely. For this reason a modified version of the software is adopted because only the region closed to the 3D space is analysed precisely (near zone) while for the rest of the tunnel that is not directly affected by the fire itself a simple 1D network is used. This model will be adopted in order to analyse how the ventilation system of Frejus tunnel has to react in order to confine the smoke trying to create a zero-velocity section closer as much as possible to the fire zone.