Commentary On 16.3

It is not valid to interpolate results from tests that are not identical except as exempted in 5.1. Any change in fastening pattern or substrate or differing end/ intermediate span ratios can have a marked effect on the result. Likewise, the practice of making the insertion of every screw 'perfect' for test purposes is an invalid practice because this does not happen in real life. Overdriving of screws, off-centring or driving a fastener that is not at right angles to the sheeting can make a significant difference to the failure load.

As the purpose of the tests is to determine the effects of the wind uplift pressure or a point load, it is essential that the arrangement of intermediate and end spans of the test specimen will provide similar results to that encountered by the actual building installation. A single span or even a two span test arrangement will not accurately simulate the reaction forces and bending moments at the purlin fastening positions, with a consequence that the actual real-world performance cannot be accurately determined or interpolated.

Calculation of performance is considered invalid because it assumes the section properties remain static under load.

Although in the past much of the testing of N.Z. roof and wall cladding has been done with a three equal span arrangement, it evidenced awareness that end spans should be approximately 2/3 of intermediate spans, which means that a three span test which consists of two end spans has no true intermediate span. Because the reaction forces are not the same it is not possible to test a true intermediate span without using a four span test arrangement.

A further complicating factor in the determination of performance loads is the use of varying or different fixing patterns. Many pierce-fixed roof cladding profiles are not fixed on every rib which gives rise to variability in the serviceability results, depending whether the fixings on the missed ribs are co-linear or staggered.

When using a four span test configuration, there is only one fully loaded central purlin. If the penultimate purlins have a staggered fixing pattern in relation to the central purlin, the rib that is fastened at the central purlin will fail at approximately 10 -15% less load than had the fastening pattern be fully linear. Linear fastening is where the same rib is fastened at each purlin and the adjoining ribs are not fixed, except at the ends. Where a staggered fixing pattern is to be tested a minimum of five spans must be used.

The end span is reduced to approximately 66% of the intermediate span because at one end it is simply supported. This proportion coincidentally approximates the increase in the peripheral wind load at the proximity of the end spans and is also often the area that suffers the highest point loads caused by people accessing the roof structure.

A plot of log P v log L has straight line fit with R2 better than 0.95 will provide a means of determining that the extrapolation is statistically reliable.

The most common serviceability limit state failure mode is permanent local deformation, which commonly precedes fracture or buckling.

Current research indicates that small changes in load can significantly affect the fatigue performance of the cladding.