Energy Education Science And Technology Part A-Energy Science And Research, cilt.30, sa.2, ss.805-816, 2013 (SCI-Expanded)
This paper is experimentally investigated, and developed a numerical model for predicting, wind
loads on roofs with various geometries. Using computational fluid dynamics simulation, several
turbulence models were compared (k-ε, RNG, Grimaji, Baldwin-Lomax zero-equation, New k-ε,
Shi-Zhu-Lumley). The Baldwin-Lomax turbulence model resulted in the most accurate results, and
thus was adopted for further analyses. Pressure distributions on gabled roofs of different slopes
(α=10o
, α=20o
, and α=30o
) for various wind speeds and directions were measured in wind-tunnel
experiments, and also predicted using the model. Comparing the values from the experiments to
those estimated by the model revealed that the model sufficiently predicted both the local and
average pressures formed on roofs for various wind velocities. There were, however, some minor
discrepancies between measured and predicted pressure coefficients, mostly at roof corner points
and back surfaces. The present work not only provides detailed information on wind load for roofs
of specific geometries, but fills a void in current wind load standards and codes of practice, which
lack sufficient data on middle regions of roofs with high slope angles. Moreover, the present study
helps lay the foundation for eventually replacing expensive and time-consuming wind-tunnel tests
with a model that accurately determines wind loads before roof production and installation,
therefore minimizing potential roof damage and losses of life and property due to roof collapse.
However, the model would benefit from further refinement and verification to improve its
capability.