(vc_atmospheric_boundary_layer)= # Atmospheric Boundary Layer ## Why this case matters The atmospheric boundary layer is the defining flow of computational wind engineering: every wind-loading, pedestrian-comfort and microclimate study begins by reproducing it. It is also the most demanding inflow to generate, because a faithful ABL must match the mean logarithmic profile {eq}`vc_atmospheric_vel`, the turbulence-intensity profile {eq}`vc_turbulence_intensity` and the target spectra at once, all with the correct dependence on the terrain roughness length $z_{0}$. This case validates that whole chain against the standards and wind-tunnel practice used in industry: the Eurocode 1 profiles {footcite:t}`en2005eurocode`, the Counihan-method tunnel setup of {footcite:t}`wittwer2013statistical`, and the roughness-element sizing of {footcite:t}`ricci2018towards`. It is the bridge between the solver's inlet-turbulence and wall-model machinery and a real wind-engineering result. The atmospheric boundary layer is localized on the lowest part of the atmosphere, where the wind flow is highly influenced by the proximity to the ground, which generates a turbulent flow, as illustrated below: ```{figure} /_static/img/solver/validation/cases/atm_boundary_layer.svg --- width: 70% align: center --- ``` When assessing the wind load in constructions it's necessary to consider an atmospheric turbulent flow condition. The setup of a numerical model that reproduces this kind of flow mimics an experimental wind-tunnel. ```{note} The wind-tunnel may reproduce some aspects of a atmospheric flow, however it does not give a completely accurate representation. Contrary to the atmospheric flow, the wind tunnel has well defined boundary conditions. Also, it is not able to represent the Coriolis forces or achieve the high Reynolds number present on the atmosphere. ``` The height and geografical density of obstacles in the terrain, such as trees and constructions, affect directly the caracteristics of the local athmospheric boundary layer. In order to account for the different profiles caractetistically found in certain region, such as lakes, rural areas, urban areas, etc., technical standards for the assessment of forces caused by the wind over buildings present a set of categories or exposures: | | | | | | -------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | --------------- | -------------------------- | | **Terrain Category** | **Description** | $z_{0}$ **(m)** | $z_{\mathrm{min}}$ **(m)** | | 0 | Sea or coastal area exposed to the open sea | 0.003 | 1 | | I | Lakes or flat and horizontal are with negligible vegetation and without obstacles | 0.01 | 1 | | II | Area with low vegetation such as grass and isolated obstacles (tress, buildings) with separations of at least 20 obstacle heights | 0.05 | 2 | | III | Area with regular cover of vegetation or buildings or with isolated obstacles with separations of maximum 20 obstacle heights (such as villages, suburban terrain, permanent forest) | 0.3 | 5 | | IV | Area in which at least 15% of the surface is covered with buildings and their average height exceedes 15 m | 1.0 | 10 | Where $z_{0}$ is the roughness length, the main constant defining the shape of the logarithmic profile applied, and $z_{\mathrm{min}}$ is the minimum height above which the logarithmic law can be applied to estimate the velocity. Altough $z_{0}$ cannot be interpreted as a direct measure of obstacles mean height, some empirical relations can approximatelly relate both notions, as it will be seen further. From {footcite:t}`en2005eurocode`, the time-averaged velocity profile is calculated with: ```{math} --- label: vc_atmospheric_vel --- \bar{u}\left(z\right)=c_{r}\left(z\right)c_{0}\left(z\right)U_{\infty} ``` where $c_{0}$ is the topography factor, usually assumed as 1.0, and $c_{r}$ is the roughness factor, calculated as: ```{math} --- label: vc_roughness_coefficient --- \begin{aligned} c_{r}\left(z\right)=k_{r}\mathrm{ln}\left(\frac{z}{z_{0}}\right) \quad \mathrm{for} \quad &z_{\mathrm{min}}\leq z \leq z_{\mathrm{max}} \\ c_{r}\left(z\right)= c_{r}\left(z_{\mathrm{min}}\right) \quad \mathrm{for} \quad &z \leq z_{\mathrm{min}} \end{aligned} ``` where $z_{\mathrm{max}}=200$ m, and $k_{r}=0.19\left(z_{0}/z_{0,II}\right)^{0.07}$. Another quantity of great importance is the turbulence intensity $I_{v}$, whose reference standard values are calculated as: ```{math} --- label: vc_turbulence_intensity --- I_{v}\left(z\right)=\frac{k_{l}}{c_{0}\left(z\right)\mathrm{ln}\left(z/z_{0}\right)} ``` In order to reproduce such profiles of mean velocity and turbulence intensity, experimental wind tunnels must position a pattern of obstacles so as to simulate an equivalent terrain's roughness condition. A general wind tunnel setup is illustrated below: ```{figure} /_static/img/solver/validation/cases/wind_tunnel_example.svg --- width: 70% align: center --- ``` The building in which the wind load will be measured is positioned in a rotative table. To build the correct wind profile, the fan is usually followed by a grid to produce a turbulent inflow. The illustrated setup follow {footcite:t}`wittwer2013statistical`, which uses a castellated barrier right at beginning of the flow, followed by elliptical vortex generators and a sequence of obstacles. This is the Counihan method for the simulation of atmospheric boundary layer. The obstacles are built following the proposal from {footcite:t}`ricci2018towards` in which the following relation is applied: ```{math} --- label: vc_obstacles --- z_{0}=0.5h\frac{A_{r}}{A_{t}} ``` in which, $h$ is the obstacles height, $A_{r}$ is the roughness element area perpendicular to wind direction, and $A_{t}$ is the ground area per roughness element. The velocity profiles normalized to 10 meters and the turbulence intensity for the mentioned terrain categories is shown in the plot below:
/_static/img/solver/validation/cases/velocity_categories.svg
Velocities for each category
/_static/img/solver/validation/cases/intensity_categories.svg
Turbulence intensity for each category
## Synthetic eddy inlet (SEM) The inflow turbulence is generated with the synthetic eddy method (see the {ref}`bc_synthetic_eddy_method` theory section), which reproduces, per inlet node, the target mean velocity profile and Reynolds stress tensor. The per-category profiles are derived from the Eurocode 1 targets above (Eqs. {eq}`vc_atmospheric_vel`-{eq}`vc_turbulence_intensity` and the Annex B length scale). Eurocode 1 specifies only the longitudinal statistics; the lateral (`v`) and vertical (`w`) intensities and length scales come from established wind-engineering heuristics (ESDU 85020/86010; {footcite:t}`solari2001probabilistic`; Counihan 1975; Dyrbye & Hansen 1997): $I_v = 0.75\,I_u$, $I_w = 0.50\,I_u$, ${}^{x}L_v = 0.25\,{}^{x}L_u$, ${}^{x}L_w = 0.10\,{}^{x}L_u$. The ratios are explicit, tunable constants in `eurocode.py`. ```{toctree} --- hidden: --- 01_atm_BL_sem.ipynb ``` ```{footbibliography} ```