Signed in as:
filler@godaddy.com
CFD Workflow
MODEL SYSTEM
The first step is to gather information necessary to create a virtual representation of a physical system. In general, this information includes the dimensions and locations of all major structures that affect wind flows within the regions under consideration. A solid body representation of the system is typically generated using Solidworks 3D CAD design software.In this example, airflow over the example open storage yard is considered. The storage yard includes four storage piles, four local buildings, an upwind greenbelt, and a surrounding wind fence.
MESH
The mesh consists of 8 million tetrahedron elements. To ensure a quality solution, a denser mesh is used within the local region surrounding the stockyard, including the greenbelt, where smaller scale flow effects are more likely to arise. The mesh also has an increased resolution near all solid surfaces to ensure adequate resolution of turbulent wall-effects. This boundary layer mesh, located in close proximity to all solid surfaces within the system, consists of a region 10 elements ‘thick.’ The element spacing in the boundary layer varies according to the element face size at the corresponding fluid-solid interface.
TURBULENCE MODEL
To compute the airflow in this system, Fluent solves the Reynolds averaged Navier-Stokes (RANS) equations using an eddy-viscosity assumption. Of the turbulence closure models available within Fluent, the realizable k-epsilon model has been found to be the most accurate for modeling the aerodynamics of atmospheric flows over windbreaks and similar structures [Bourdin and Wilson, 2008].
RESULTS (VERIFICATION)
Iterative convergence is discussed in the previous section. Most of these stockyard CFD simulations require at least 2000-3000 iterations to fully converge, and in some cases, up to 5000-6000 iterative steps are necessary. It is important to monitor the error to ensure that the results are representative of a fully converged simulation. Grid independence is another important issue that must be addressed in these studies. The results shown here are for the initial mesh of 8 million elements. The standard approach to show grid independence is to increase the number of grid points from 8 million, to say 12 million, and again to say 16 million, and plot the results as a function of increasing mesh resolution. The model will approach grid independence as the mesh resolution is increased, and the final mesh is chosen once you either (1) find a situation where the results cease to change appreciably or (2) reach the computational limits of your finite computational resources.
RESULTS (VALIDATION)
A CFD model is ‘validated’ when it is shown to accurately represent experimental measurements. For these large scale atmospheric flows over industrial stockyards there is typically a lack of empirical data. However, it must be noted that previous researchers have taken extensive steps to obtain experimental data and perform comparative studies of many turbulence models. The models which most accurately represent the experimental data were noted and used in this work. In that sense, the model has been validated against other situations, and this gives us more confidence as we extend it to these stockyard scenarios.
FUGITIVE DUST FATE AND TRANSPORT
Once the flow field is established, particles are released from the surfaces of the piles. Fluent’s discrete-phase modeling is used to model the dispersion of these particles due to the effects of the flow field, the turbulence, gravity, and the interaction of the particles with the ground, buildings, windfences, and greenbelts. The reactions of individual particles are determined by solving the equation of motion for each particle in the system. In most studies, the primary results include analysis of the fate and transport of aerosolized dust particles to investigate the aerodynamic sheltering effects of various windfence configurations.
WIND SHEAR AT EMISSION SURFACES
The entrainment of dust particles into the air is correlated with the shear stress at the air-ground interface. Dust is entrained into the air when this wind shear reaches a critical value. Previous studies have focused on characterizing this threshold for a variety of dust materials. Contour plots of surface stresses, as well as near ground velocities, can readily be generated to examine probable locations that exceed this threshold. The efficacy of wind fences and greenbelts at reducing the critical stress and/or threshold velocity can be directly computed using CFD.