Pressures on a cubic building-Part 1: Full-scale results

Abstract

Pressures on the vertical and horizontal centrelines of the Silsoe 6m cube are presented. The full-scale data has been processed in 12. min non-overlapping record blocks of cube surface tap pressure together with the reference upstream approach flow measured at cube height. Since in full-scale one has to accept the wind conditions provided by nature, each block of data is unique in regards to even the mean conditions and there is no opportunity to exactly replicate a test multiple times, as might be done in a wind tunnel. As a result the data are processed in a manner that makes use of the large number of blocks recorded. Pressure coefficients are determined by normalising with the wind dynamic pressure at the reference position. The mean, standard deviation, maximum and minimum pressure coefficients are determined by normalising the measured pressures by the mean, standard deviation and maximum wind dynamic pressure, respectively. It is concluded that using this normalising process helps to minimise the uncertainty of the results. A least squares optimisation method is used to fit a truncated Fourier series to the pressure coefficient data over the 360° of flow direction. The differences between the data points and the fitted curves are used to assess the uncertainty associated with each type of coefficient. It is shown that typically 90% of the normalised data lies within a band of ±0.1 for the mean pressure coefficient and ±18% for the standard deviation coefficient. However with the maximum and minimum pressure coefficients the uncertainty is higher and asymmetric, and is best modelled as a combination of a fixed amount plus a percentage of the coefficient. The highest uncertainty noted was for the minimum coefficient where 90% of the data lies between about +0.1-16% of the coefficient and -0.1+25% of the coefficient. It is shown that all of the pressure coefficients determined form a consistent set, which exhibit sensible variations with both position and wind direction. It is believed that this set of data offers opportunities for the assessment of wind tunnel and CFD models, which has not been previously possible.