Simulation of Fluid Dynamics around a Cubical Building

This study describes a validation of NRL’s FAST3D-CT (Contaminant Transport) model directly relevant to the accurate simulation of the transport and dispersion of contaminants external to buildings. The study is of flow over a surface-mounted cube. Fluid dynamics can become arbitrarily complex in even simple flow geometries when turbulence is present. Therefore, since the cube is a single parameter representation (i.e. size) that approximates many buildings, it is an excellent first test problem. Of particular interest is understanding how to realistically simulate atmospheric wind conditions in computational models, so that the resulting flow patterns around buildings are accurately described. To this end we investigate the form of the inflow velocity on flow and dispersion patterns around a cubical building. Computational results are compared with wind tunnel and water tank data published by a number of authors. Because an essential feature of validation is a thorough understanding of the inflow conditions, wind tunnel and water tank experiments are really the only place where satisfactory validation, in the usual sense, can be performed. The extensibility of the results depends on the simulation model’s ability to represent more complex geometries easily and faithfully.

In the present investigation of flow over a surface-mounted cube, we use the model FAST3D. FAST3D is flow solver for three-dimensional, time-dependent, compressible reactive flow problems. The underlying fluid dynamics algorithm is the Flux-Corrected Transport (FCT) algorithm, which is a high resolution, direction split, monotone, conservative, positivity preserving algorithm. The turbulence is modeled through the Large Eddy Simulation method MILES, in which subgrid effects are accounted for implicitly by the non-linear flux limiting of the algorithm. In the study, a particular configuration was chosen as a baseline case about which parameter excursions were computed to understand the important physical and numerical factors influencing the evolution of the fluid dynamics. A key aspect of the study was to investigate contaminant dispersion patterns. In the present work we therefore consider the effect of a contaminant source located a short distance downstream of the cubical building, and compare our numerical results with available experimental data. A grid resolution study was performed to assess the effect of resolution on the features of the flow field, particularly with respect to vortex shedding and source dispersion. To further investigate the effect of ambient wind conditions around buildings, the effect of non-uniform, boundary-layer-type inflow velocity profiles on the flow structure was also investigated.