The aim of this test program is to ascertain the effects of high wind speeds on Plastex Crossgrip roof walkway mats; in particular to establish at what speed they are likely to lift away from the roof surface. These lift off speeds can then be compared with the wind conditions that could be experienced at potential installation sites.
The Beaufort Wind Scale
In 1805 Admiral Sir Francis Beaufort developed a scale to help sailors estimate the winds via visual observations. The Beaufort Scale extends from force 0 to force 12 and still used today to estimate wind strengths on land and at sea. The inclusion of this information in this report helps to relate the wind speeds at which the mats were tested to real weather conditions.
The testing was carried out using the open return wind tunnel in the aerodynamics laboratory of the School of Aerospace Automotive and Design Engineering at the University of Hertfordshire. This tunnel has a working section of 480 mm diameter. A fan driven by an electric motor provides close control of air velocity up to a maximum of approximately 100 mph (45 m/s).
The air velocity was measured using calibrated pressure tappings built into the tunnel. An independent check of the velocity measurement was made using a hand held pitot-static tube and an Airflow PVM100 micromanometer. The results of this velocity check showed a variation between the two measured wind speeds of less than 2% over the range 25mph – 100mph.
In order to fit within the tunnel working section the mats were cut to a length of 1222mm (4 feet). For the testing of mats supported by the bars running parallel to the air flow the test sample width was 475mm. For the testing of mats supported by the bars running perpendicular to the air flow the test sample width was 470mm. This small difference is due to the mats being cut adjacent to a bar.
The ground board supporting the mats was covered with standard roofing felt. However, to ascertain if surface finish was significant, some testing was performed on a smooth-surfaced plywood ground board.
The mat was placed in the tunnel in two orientations; with the support bars running parallel and perpendicular to the air flow. Testing was also performed with the leading edge of the mat raised slightly from the ground board, by wire support from the upper frame of the tunnel, to simulate possible installation scenarios. The mat was also tested in a partly rolled back arrangement. The wind speed was gradually increased until the mat lifted from the surface. This speed was recorded and some video footage was taken of the event. A selection of the test arrangements were repeated to confirm the values for the mat lift-off speeds.
The results showed a consistent trend with good repeatability.
For the mat placed on either a flat standard roofing felt surface, with a natural roughness or a flat smooth surface, with the solid surface of the support bars exposed to the wind direction the mat did not slide or lift for the full wind speed range of the tunnel, which was just in excess of 100mph. However, at 100 mph the leading edge of the mat exhibited some small, (a few mm) vertical oscillations.
For the mat placed on a flat smooth surface with the channels formed by the support bars parallel to the wind direction the mat did not slide or lift for the full wind speed range of the tunnel which was just in excess of 100mph.
For the mat placed on a flat standard roofing felt surface, with a natural roughness, with the channels formed by the support bars parallel to the wind direction the lift-off point was very sudden and preceded only by a slight twitching and lifting of the mat leading edge. This result was confirmed by repeat testing. It is likely that small flow perturbations and boundary layer development caused by the surface roughness of the ground board are sufficient to cause small amounts of lifting of the mat leading edge into the incident airflow which then lead very rapidly to lift-off of the whole mat.
For all arrangements where any part of the leading edge of the mat was lifted then the speed at which total lift-off of the mat occurred was generally a function of the amount of initial leading edge distortion away from the ground board.
For the arrangement where the leading edge of the mat was lifted locally, the whole of the leading edge lifted with increasing wind speed and attained a state of lift – weight balance prior to complete mat lift- off. The lift off speed was similar for both the roofing felt and smooth ground board surfaces.
For the partially rolled up mat arrangement the lift-off speed was relatively low with the gradual lift process preceded by a rocking of the rolled mat.