Development and calibration of a 1D thermo-fluid dynamic model of ventilation in tunnels

In complex, large civil infrastructures where ventilation has a crucial role for the safety of users in both normal operation and hazardous scenarios, the correct prediction of flow and heat transfer parameters is of fundamental importance. While full 3D simulation is applicable only to a limited extent, and the resort to 1D modeling is a common practice in both design and evaluation phases, the limitation of such models lies in the choice of transfer parameters, such as friction loss coefficients and heat transfer coefficients. In this work, we present an original approach based on the Finite Volume integration of the 1D flow and energy equations on a network of ducts, representing the ventilation system in the 11.6 km long Mont Blanc Tunnel with a spatial resolution of 10 m. The calibration of a set of friction loss coefficients against a rich experimental dataset collected throughout a dedicated set of in situ tests is of particular concern here, as it is carried out by means of genetic optimization algorithms. Predictions of the flow field are in remarkable agreement with the experimental data, with an overall RMS error of ± 0.42 m/s. Further refinements and possible parameter choices are also discussed.