2 results
Search Results
Now showing 1 - 2 of 2
Item Pixel-based Wake Interaction and Power Estimation for a Wind Farm with Irregular Boundary(The Eurographics Association, 2018) Ohlsen, G.; Ruiz-Salguero, O.; Full, T.; Acosta, D.; Livesu, Marco and Pintore, Gianni and Signoroni, AlbertoIn the domain of generation of wind turbine energy, it is central to correctly estimate the interactions among the various turbines in a wind turbine farm. The spatial super-position of turbine wind wakes determines the wind conditions that each turbine in the farm is exposed to and its power output. The current state of the art represents the turbine wakes as a 2D real-valued polygonal trapezoid. The interactions among wakes imply Boolean operations among many trapezoids, producing an intractable fragmentation of the wake intersection and domain regions. The plan (2D) view of the terrain with this wake polygon fragmentation is then used to estimate the effective wind that each turbine receives. This calculation leads to cumbersome computation, which is even more impractical if 3D representations of the terrain, wakes and wind are needed. In response to these limitations, this manuscript presents a method in which the 2D turbine wakes are located on a terrain with holes and exclusion zones bounded by 2D polygons, considering wind direction and turbine array basic specifications. Then, a discretized or pixel approximation of the terrain and wake superposition is calculated using discrete levels of the turbine velocity deficits. This process allows a practical approximation of the power output of each turbine and of the full turbine set. The wake interaction and terrain boundaries are then texture - mapped onto the 3D representation of the terrain, for visualization purposes. As an application, an example of a complex polygonal terrain turbine farm is optimized for maximal power output. This discrete, image - based calculation is particularly convenient in a circumstance in which graphics hardware and GPU processors become increasingly available and efficient, in laptop and mobile devices. This investigation opens research opportunities in mixtures of turbine types, 3D modeling of wind / terrain interaction, and accelerated calculation and visualization with GPU hardware.Item Direct Scalar Field - to - Truss Representation and Stress Simulation of Open Pore Domains(The Eurographics Association, 2018) Munoz, J. M.; Ruiz-Salguero, O.; Montoya-Zapata, D.; Cortes, C.; Cadavid, C.; Livesu, Marco and Pintore, Gianni and Signoroni, AlbertoIn the domain of lattice and porous material geometric modeling, the problem of data size is central. When using full 3D manifold Boundary Representations (BRep), even the smallest domains engender staggering amounts of 3D finite elements. A partial solution has been implemented, which represents slender solid neighborhoods with non-manifold Boolean union of 1- manifolds (curves) and/or 2-manifolds (surfaces), added with thickness information, called 1.5D and 2.5D models, respectively. Automatic applications of these techniques requires the estimation of the medial axis of the porous media, to produce a truss or frame FEA. Previous works require explicit synthesis of the skin of the porous domain. This manuscript presents an alternative in which the medial axis and thus the 1.5D (truss) representation of the porous domain is directly obtained from the scalar field (i.e., Computer Tomography -CT-) of the domain, thus avoiding the explicit calculation of the domain skin. The manuscript also presents the noise removal and linearization of the medial axis data, to obtain the skeleton truss graph (including bar radii), that represents the porous domain. Shear and tension load simulations are conducted with the Truss model, showing that the generated model can be used in FEA software. Future work is required in extending this concept to lattice materials, where the medial axis includes surfaces and not only curves, as in this manuscript.