Seminar - Matt Mason - Physical simulation of thunderstorm winds for testing structural loads
Wednesday, May 12, 2004, 1.10 - 1.50 pm
Civil Engineering Lecture Room 3
In many parts of the world, thunderstorm downbursts are believed to be responsible for much of the recorded extreme gust wind speeds used for structural engineering design codes. For determining velocity profiles with height, wind loading codes such as AS1170.2:2002 fit a simple log based profile to extreme 10 m gust wind speed. The log based profile is used to model the earth's near surface, atmospheric boundary layer (ABL), and assumes horizontal velocity flow. The ABL assumption is good for synoptic winds, but fails to represent a downburst's vertical profile adequately, and does not account for the highly rotational flow that is produced.
Work has been carried out using a pulsed impinging jet to physically model (at a reduced scale ~1/3000) characteristics of a thunderstorm downburst so that a representative vertical and horizontal profile can be determined, similar to the way the log profile represents the ABL. Both an outlet situated aperture and an outlet situated membrane have been used to produce the pulsed flow at the leading edge of steady impinging jet flow. The main aim of the pulsed jet is to simulate the impinging roll-vortex which has been observed at full-scale to lead a downburst's flow, and also study the vortex's movement along a flat testing surface. Theoretically the leading roll-vortex should contain the peak velocity, and experimentally this was shown to be the case for both aperture and membrane pulsed jets. As expected velocity time histories suggested a differing vertical profile to that assumed for the ABL, and flow visualization techniques have shown the abrupt changes in direction that would be expected from vortical flow, but is not accounted for in design codes. Comparisons to full-scale events have also been made for individual point time histories with positive results. Due however to a lack of low level full-scale data only preliminary comparisons could be made.