Experimental and Numerical Investigation of Wind Flow Around Gable Roof Low-Rise Buildings with Different Geometries

The wind flow characteristics around low-rise buildings with different geometries are investigated experimentally and numerically in this study. Computational Fluid Dynamics (CFD) simulations are conducted with simulation methods of Reynolds-averaged NavierStokes simulation (RANS) and large eddy simulation (LES). Scale-down models are tested in wind tunnel and the acquired results are used to validate simulation data. The experimental results are expected to verify the provided provisions in standards and supplement the cases that are not included in standards. With the rapid development of CFD, it is obtaining more acceptance over the last few decades. Simulations results from RANS with different turbulence models and LES are compared and validated by wind tunnel tests to evaluate the performance of CFD and determine a suitable simulation method for different flow cases. Besides, the critical flow regions around buildings with different geometries are identified. Both the significantly higher positive pressure and suction pressure generated in critical regions can cause partial failures of a building more easily than the other parts, and should arouse enough attention from wind engineers. Results show that different opening configurations, specifically, an enclosed building, a building with a windward opening, a building with a windward opening and a sidewall opening, and a building with a windward opening and two sidewall openings, have little influence on the external pressures. But these different opening configurations can result in large difference on the internal pressures and cause dramatically different net pressures acting on a building. Different roof angles can generate totally varied external pressure distributions when the wind direction is perpendicular to the roof ridge, and this difference is almost negligible when the wind direction is parallel to the roof ridge. When the wind direction is oblique to the building, there is a general tendency for high suctions to reduce significantly when the roof pitch becomes steeper. This is because vortices with higher turbulence are always generated on the roof due the sharp edges of a lower roof pitch. By comparing the RANS simulation results from different turbulence models, 𝑘 − 𝜔 𝑆𝑆𝑇 turbulence model has the best performance in pressure prediction compared to the other models. However, steady state RANS simulation method shows relatively large discrepancies in the prediction of roof corner vortices. Large eddy simulation has improved accuracy over RANS in the pressure prediction of corner vortices when a building is under oblique wind directions. Both the absolute values of pressure coefficient and the pressure distributions on the roof can be better simulated by LES. The reason is attributed to the high turbulence generated in this flow problem, where a transient simulation method could capture high intensity turbulence better than a steady state simulation method. Furthermore, the effects of balconies, roof overhangs and roof shapes, are studied to determine the impact of these subtle changes. The existence of balcony could effectively reduce the pressure coefficients on the windward wall, and building with roof overhang suffers higher suction pressure near the leading edge, which makes the roof structure more vulnerable than building without overhang. The different roof shapes, including flat roof, gable roof and round roof, generate similar flow patterns around building and similar pressure coefficients on the building surfaces. When the wind direction is perpendicular to the roof ridge, more severe suction pressures can happen on the roof when the roof shape changes in the order of flat, gable and round. The change of roof shape has little influence on the magnitude of suction pressures brought by roof corner vortices at a wind direction of 60° but can affect the pressure distribution patterns.