Results are presented from a series of wind tunnel tests to measure the cav
ity pressure dynamics of a flexible roofed building with a single dominant
opening. Tests were conducted in a simulated atmospheric boundary layer, re
presentative of an open-country terrain, for the model configured with five
different roof tensions, three internal volumes, two windspeeds and five o
pening azimuth angles. The ratio of air-to-building bulk moduli was used as
a measure of roof tension and buildings with 0.02 < K-a/K-b < 7.5 were tes
ted. It was found that the cavity scaling parameter (Eq. (1)) provided a co
nvenient means of collapsing the unsteady cavity pressure data for those co
nfigurations where significant resonance occurred. A flow component into th
e opening was necessary to sustain large resonant oscillations and this occ
urred with an opening on both windward and leeward sides of the building. I
ncreased flexibility of the roof membrane reduced the value of the Helmholt
z frequency and also reduced the magnitude of the resonant response. (C) 19
99 Elsevier Science Ltd. All rights reserved.