Multiscale modeling of transient flows from fire and ventilation in long tunnels

This paper applies a transient multiscale approach to model ventilation flows and fires in a tunnel domain. The multiscale model couples dynamically a Computational Fluid Dynamics (CFD) solver with a simple 1D network model, allowing for a more rational use of the computational resources without loss of accuracy. The 1D network models tunnel regions where the flow is fully developed (far field), while detailed CFD models regions where flow conditions require 3D resolution (near field). The paper describes both numerical models and gives emphasis to the discussion of the coupling algorithm and the control of the numerical error.Compared to full CFD, the multiscale model provides a reduction of the required computing time by 40 times without significant loss of accuracy. The methodology has been applied to study the transient flow interaction between a growing fire and a ramping-up ventilation system in a modern tunnel of 7 m diameter section and 1.2 km in length. Different ventilation scenarios are investigated to provide the timing to reach the critical velocity at the seat of the fire, and to remove the upstream back layering. The results allow for simultaneous optimization of the ventilation and detection systems. The multiscale methodology represents the most feasible tool to conduct accurate simulations in long tunnel domains, when the limitation of the computational cost becomes too restrictive.