Background to wind load testing

It is well known and accepted that New Zealand suffers from frequent and numerous gales, which sometimes approach the wind speed of the cyclonic areas of Australia and the Pacific Islands. This situation has been assisted by the publication of AS/NZS 1170, which places the emphasis on specific location rather than wind speed, and factorising a basic speed dependent on terrain and topography.

Wind is a dynamic force and the method of testing sheet roof and wall cladding solely by a static load is not logical. Any test procedure should as closely as possible resemble the loading conditions likely to be experienced by the cladding in the field during its lifetime. This anomaly has come about historically because it was expedient to do so and the cost of cyclic testing, as the alternative to static testing, was considered to be unjustifiable. The validity of static testing results is questioned, particularly when cyclic testing of roof and wall cladding has shown that fatiguing can occur at load level much lower than the static test results. This is particularly true of high strength steel, which has been and continues to be a common material used for domestic and industrial roofs in New Zealand.

For these reasons the New Zealand Roofing Industry has established a cyclic regime suitable for verifying the static testing of roof and wall cladding in New Zealand. Factoring static test loads alone has been considered and discarded as being inaccurate and invalid.

Results from many tests using high strength steel points to a value at which no damage will occur even up to 10,000 cycles but a small increase in value will quickly initiate crack propagation within a 1000 cycles. It is for this reason that cyclic tests have been increased incrementally. Once the target load has been passed successfully the test can be continued incrementally until failure.

The point at which permanent deformation occurs in HS Steel is very close in value to the point when cracking will be initiated and eventually fail over a prolonged cyclic regime. Because of the consistency of test results obtained, there appears to be no justification to test to 10,000 cycles as required in other cyclic test regimes. The use of cyclic testing is intended to be used as a check for the static test as the frequency of cyclic test checking can be adjusted for the material. i.e. low-strength steels are unlikely to suffer fatiguing at their lower levels of performance.

The importance of cyclic testing is to provide a performance value for the area of cladding serviced by a single fastener, (the Tributary Area) below which fatiguing is unlikely to occur. Once crack initiation has begun it is inevitable that failure will occur under repeated load.

The fastening pattern is a major determining factor where failure due to fatigue is determined not only by the number of fasteners per m2, but also by the washer size as can be seen by the greatly increased performance when using load spreading washers.

Any failure is likely to be subjective, as it is with the static test, but any visible cracking or permanent deformation around the fastener in position is deemed to be failure. This means that both the static and the cyclic tests can never be fully automated and will require a physical examination and assessment.

Because this standard is material non-specific, it requires subjective judgment as to (say) if the fibreglass reinforcement of GRP sheeting has been permanently deformed.

Where clipped sheeting (secret-fix) is used the design value for the tributary area can be a lesser value than for pierce fixed sheeting. Because this value is dependent on the profile shape and clip design, the tributary value leading to the target load, must be established for each profile.

It is common practice in NZ to use plastic sheeting in conjunction with metal sheeting at the same purlin spacings. For this reason plastic sheeting should also be included in any test programme.

The tributary load (TL) is calculated by multiplying the area serviced by the fastener by the test load in kPa. Because fastening patterns are not always symmetrical the maximum mean distance between fasteners is used.

Some TL values have been given in the NZMRM Code of Practice for specific thicknesses, profiles of steel, with and without using load spreading washers. Recent tests have confirmed these values however different thicknesses, grades of material and different thicknesses, grades of material, metal and shapes of load spreading washers will give very different results. For this reason their TL values must be established by the method prescribed.

All static tests must be subject to checking that the tributary area of the purlin spacing and the fastening spacing fit within the 'no cracking' criteria established and verified by cyclic loading.

The static test load is applied incrementally to provide opportunity for the observation of the performance of the profile under load. A 'feel' for this performance is easily obtained after conducting a number of tests and these can be used as a practical learning tool for roofers in training.

The use of load spreading washers has a major bearing on the performance of sheeting under negative loading which is mainly due to the increased area over which the load is shared. The most commonly used screw fasteners used in NZ are 12# & 14# fasteners and both sizes of Type 17 for timber and self-drilling for steel, have the same screw head diameter. This means that there is no advantage in using a larger screw for 'pull-over'. The load spreading washer as manufactured in New Zealand follows the profile of the sheeting and provides an increase in profile performance of up to 100% as well as providing a seal for the fastener. Because the increase of load of up to this amount applies at the periphery of a building, by using load spreading washers, the additional load can be taken by the sheeting without changing the purlin spacing or fastener frequencies.

Calculation does not take into account the changing section properties of the profiled sheeting with increasing load and nor can it allow for the interaction and support provided by adjacent spans. These are very much related to profile and the proportionate relationship with the end and intermediate span.

The minimum number of sequential tests to provide information for a load/span graph is three. It is suggested that a cyclic test be conducted on one of each of these series per permutation. i.e. with and without load spreading washers, or staggered fixings etc.

The reason for this level of static load checking (1 in 3) is because of the varying loads placed on the fixings by different purlin spacings and span configuration. It is therefore not critical as to which span is cyclically tested but it is recommended that the median of the three spans be cyclically tested.

Because the cyclic test is a verification test only to demonstrate that the static test would not fail under repeated load, it does not require to be factored by the factors in Variability Factors.

The cyclic regime used for New Zealand does not follow the Australian or other models.

1000 cycles at 80% of the target load, cycling from 40%-80% of the target load.

1000 cycles at 90% of the target load, cycling from 45%-90% of the target load.

1000 cycles at 100% of the target load, cycling from 50%-100% of the target load.

The reason why the cycle is not dropped to zero is because gusting under storm conditions more closely resembles these sequences.

One of the reasons why deflection is regarded as a performance criterion in Australia is because of ponding. Although rainfall does not normally occur at maximum wind load, in adverse circumstances standing waves and turbulence could still give leakage at the minimum pitch allowable in Australia (1°).

It is for this reason as well as many others that NZ has adopted 3° as a minimum pitch.

Note 5 following Table C1of AS/NZS 1170 states:

Problems with visually sensed deflections are frequently dependent on the presence of a visual clue for the observer to gauge linearity. Deflections are therefore a function of the line of sight of the observer.

Although measured deflection under load is a performance indicator, performance under load is not considered a test criterion. The deflection design criteria also must be material non-specific, so the use of different metals and plastic for roof and wall cladding could impose unnecessary limitations on the use of these materials because they possess a very different Young's Modulus compared to steel.

Residual deflection is therefore set to a lower level at L/400 than would be normal for HS steel at L/1000.