The aerodynamic behavior of porous elements in external flows: modeling approaches in computational fluid dynamics based simulations

In the last decades, porous surfaces became increasingly popular in modern architecture due to their aesthetic value, the possibility to be used in energy efficient cladding solutions and their ability to maintain ventilation. However, as the size of the pores is generally one to three orders of magnitude smaller with respect to the overall structure dimensions, it is extremely difficult to study the aerodynamics of porous surfaces in wind tunnel experiments. Computational Fluid Dynamics, CFD, simulations in which the pore geometry is explicitly modelled are in no way easier than wind tunnel tests, as their computational costs are substantially unaffordable due to the large number of cells required to represent the tiny pores. In such context, an alternative approach is to use appositely defined homogenized models in order to account for the presence of the porous surfaces, usually represented by the so-called pressure-jump approach. Currently, this modeling approach has been fairly validated for internal aerodynamics, which generally provides good results, but it is not widely validated for external aerodynamics problems. In order to solve the aforementioned problem, this thesis aims at using the pressure-jump approach to simulate the porous surfaces in external flow and evaluating the performance of this approach. For this purpose, firstly, the flow through porous barriers was analyzed in detail and a new model able to predict the pressure jumps based on porosity was derived. Then, the proposed model was calibrated by comparing itself to several simulated and experimental results. Based on this calibrated model, the results obtained using pressure-jump and explicit models of the porous surfaces geometry were compared. For all cases, a systematic comparison was carried out between the two aforementioned modeling strategies. In most cases, the sensitivity of these strategies to turbulence modeling and numerical schemes were also considered.