Abstract: The present work aims to evaluate the influence of porosity of artificial barriers on wind velocity reduction using numerical simulations, testing, at the same time, turbulence RANS (Reynolds Averaged Navier-Stokes) models to obtain better predictions of the flow field. In order to support the analysis held in the present work, experimental studies by Dong et al. (2010) were chosen to validate the numerical simulations carried out. Dong et al. (2010) investigated the influence of the porosity of artificial barriers on the wind flow and performed detailed measurements of the velocity and turbulence fields downstream the artificial barrier model through Particle Image Velocimetry (PIV) technique. In the following sections the experimental work is described, as well as the physical and mathematical modeling of the artificial barriers and the studied cases. Artificial barriers are generally treated as a porous medium. The physical effect of a porous medium on the flow is a pressure drop. The porous medium creates resistance to flow constituting a sink for momentum (Cong et al., 2011). Thus, the wind barrier promotes a decrease in the flow movement and a consequent reduction on wind velocity downstream the barrier providing a shelter effect in this region. The resistance generated by the porous barrier can be take into account in the momentum equation by means of the inertial resistance in a source term. The flow resistance is expressed by the pressure loss coefficient or inertial resistance coefficient and can be described as a function of the porosity of the barrier. Several equations are used in the literature to describe the relationship between the porosity and inertial resistance coefficient, however few authors have evaluated the effects of these formulations on the behavior of wind flow, velocity reduction and turbulence. In the present work, the inertial resistance coefficient was modeled using different equations that relate the resistance coefficient to barrier porosity, as presented by Song et al. (2014), Xu et al. (2019) and Yeh et al. (2010). In order to evaluate the turbulent model effects on velocity reduction promoted by artificial barriers, the following two-equation Reynolds Averaged Navier-Stokes (RANS) models were chosen: the realizable κ-ε (proposed by Shih et al.,1995) and the κ-ω SST (proposed by Menter, 1994). Two-equation models are the most used, and they provide good predictions for turbulent flows in the atmospheric boundary layer. Besides, they have a lower computational cost when compared to superior RANS models. In order to perform the numerical simulation of the flow through artificial barrier, the computational domain was built to model the same geometry and atmospheric conditions that the wind tunnel experiment carried out by Dong et al. (2010). The geometry has the following dimensions: 2.80m in length, 0.50m in width and 0.40m in height. The barrier was positioned at x=0. The distance of 50H upstream and 90H downstream of the barrier was enough to assure that the inflow was not affected by the presence of the wind barrier, as well as the outlet region was at a sufficient distance for the complete flow development. Two meshes with 4.7 millions and 5.2 millions elements for simulations using realizable κ-ε (z + ≤ 11) and k − ω SST (z + ≤ 5) models, respectively, was built by triangular cell extrusion using the mesh generator of the Ansys 19.1 package. Near the surface walls, refinement using prismatic elements were produced to satisfy the requirements imposed by the models. In the region adjacent to the prism layer, the global element size (4.5×10−2 m) was the same for the two meshes in order to guarantee the same physical phenomenon representation for both turbulence models. The inertial resistance coefficient used to model the flow in porous media was calculated using different equations. That express the relationship between coefficient and porosity of the barrier accepted in the literature. The model result of simulation were compared to experimental data by Dong et al. (2007;2010) to validate the simulations cases considering the mesh, model and boundary conditions. The longitudinal horizontal velocity component (u) was dimensionless by the free stream velocity that is equal to 10m.s−1 and the height z was normalized by H, where H is the height of the wind barrier equal to 0.02m.

Keywords: Windbreak, Wind velocity, Porosity, Friction velocity, CFD, Turbulent flow