A comparison of wind-tunnel and full-scale wind pressure measurements on low-rise structures

Abstract

Comparison are made between wind pressures measured on two low-rise experimental buildings and pressures measured on wind-tunnel models of those buildings. For the experimental building at Aylesbury, U.K., comparisons are made between the full-scale pressures obtained by the Building Research Establishment and those of model tests at 1:500 scale carried out by the University of Western Ontario, Canada (U.W.O.) and at 1:50 scale by Virginia Polytechnic Institute and State University (V.P.I.S.U.). The second experimental building, constructed by V.P.I.S.U. at Price's Fork, VA, provided information on wall pressures which are compared with those obtained from a 1:24 scale model tested in the wind tunnel at V.P.I.S.U. By using pressure coefficients based on the mean velocity in the approach flow at the level of the pressure measurement, it is shown that there is little difference between mean or fluctuating pressure coefficients obtained from a model at 1:500 scale, in a carefully simulated boundary layer, and those from a model at 1:50 scale. Scaling in the latter case did not allow careful simulation of the mean velocity profile, but did provide a suitable level of turbulence and a turbulence integral scale at least as large as the largest model dimension. Use of large model-scales has the advantage that relatively small details of construction can be included. On the basis of the full-scale/model comparisons it is shown that the non-stationary character of the natural wind has a significant effect on the mean, r.m.s and peak pressure coefficients. Under non-stationary wind conditions the full-scale extreme peak coefficients may be as much as 5 times the wind-tunnel values. Local pressure coefficients can be modeled adequately for low-rise structures located on level sites (Aylesbury) with relatively uniform upstream terrain, provided the turbulence intensity and turbulence integral scale are properly simulated. For structures located on sloping sites and with complex upstream terrain (Price's Fork experimental building), the modeling of the mesoscale terrain features may be very important. The complex terrain is responsible for increased turbulence intensities of the horizontal velocity components as a results of increased low-frequency spectral energy.