11.08 Structural design

Categories: Stone Cladding

Introduction
The structural design of a stone cladding system has to take account of the required performance of the stone and of the fixings.  It should assess all possible failure modes and show that in every case the required margin exists between the expected load capacity and the expected loads. At the same time the fixings must be designed to ensure that movements of the cladding and structure can be accommodated without causing additional loads in the fixings or cladding.

The design process should evaluate the four aspects of the cladding:

• Attachment to the support wall or frame,
• Attachment to the stone panel,
• Integrity of the fixing itself,
• Integrity of the panel.

The ability of the support wall or frame to resist the loads imposed by the cladding must also be checked but is outside the scope of this Section.

As with most other structures the purpose of structural design is to ensure that the cladding and fixings are strong enough to withstand the loads to which they are subjected safely and without impairing their performance. Section 04.09 describes the design principles in greater detail.
 

Design methods
BS 8298 provides design guidance in the form of acceptable dimensions of stone panels including fixing details which can be used in many situations, however where the stone cladding does not comply with these dimensions it is necessary to prove the design by other means. Some aspects of performance may be assessed by calculation, for instance the deflection of an angle support, the tensile stress in a hanger, or the buckling load in a tie. However many aspects of stone cladding, particularly fixing design are difficult to model in calculations and must be assessed by testing. It is unlikely that all aspects of the fixing can be proven in a single test and a programme of tests or combination of calculations and tests has to be developed.

Recourse may also be made to previous use and testing of a component.  The use of anchors in concrete and blockwork is well documented and design charts exist for many products.  In this case it may not be necessary to test to prove the design.  However, site installed anchors should still be tested after installation to check workmanship and in situ strength.
 

Testing programmes
It is possible to test large full scale cladding specimens and this is common for bespoke curtain walling.  The cost of constructing and testing large specimens is justified because it is the only way to test for water penetration resistance and integrity under wind loading.  However many of the structural aspects of a rainscreen may be proven by the testing of individual panels and components, CWCT, 1998.

When testing a stone cladding system it will be necessary to test at the design load.  As the factor of safety on loading is normally different for stone, metal and concrete components there is no single proof load at which tests on complete cladding systems, or even individual panels, can be carried out.

If a large specimen comprising one or more panels of cladding is tested at a load in excess of 1.5 times the nominal load, the load is likely to produce an expected failure in the support frame, or in the fixings remote from the stone, before the anchors and stone details experience the required proof load.  Similarly, if a single stone panel is fixed as intended on site and then loaded, failure will occur in the fixings remote from the stone before the panel and stone details experience the required proof load.  For these reasons it is normally necessary to conduct separate tests for:

• Attachment of the fixing to the support frame or wall,
• Attachment to the stone (anchors, kerfs, dowels, etc.),
• Fixings remote from the stone,
• Flexure of the stone panels.

Natural stone is a variable material and when undertaking tests it is necessary to perform enough tests to give a statistically valid answer with the required degree of confidence. Typically every test is repeated ten times, more for stone with variable properties, but this is not feasible when testing a large-scale cladding panel.  A single result from a large scale test may be misleading and of less value than a series of repeated tests on a single component.

Large test panels comprising a number of stone cladding panels and fixings will be erected to the tolerances required of the test rig and the induced deviations of the stone will be ‘run of the mill’.  The tests will not then be representative of the worst possible design condition on the wall.  When testing individual components it is possible to set fixings at their maximum extension, position dowels and kerf plates close to the face of the stone and use the thinnest stone panels.  Above all it is possible to observe the failure clearly and record the dimensions of the test specimen.  With this information it may be possible at a later date to calculate stresses and engineer any required design changes.

If a large specimen has been constructed for the purpose of air and water penetration testing then it may be used to prove some aspects of the structural performance of a stone cladding system, for instance thermal movement or movement of a structural support frame.  The testing regime for a curtain wall may call for an ultimate load test, CWCT 1996.  In this test the wind loading on the specimen is increased until a failure occurs.  The test is intended to show the ultimate failure mode and may show whether failure occurs in stone panels, fixings or anchors.  Of course all individual components show a range of strengths and it may be difficult to rely on the results or even interpret them.

Tests of individual components may not be sufficient when testing some thin stone cladding systems.  If the stone panels undergo large flexural movement during wind loading or thermal and moisture movement they may impose a bending load on the fixings and the anchors, kerfs or dowels in the stone panels.  This is not generally regarded as a problem for stone panels thicker than 20mm; for thinner panels such as glass, ceramic, and structurally assisted stone veneers it should be considered.  This aspect of behaviour can only be tested by loading a complete panel or panels and the corresponding fixing system.
 

Calculation
Where the stress distribution in a component can be established by calculation, the design of the component may be based on calculation of the load bearing capacity of the component.  BS 8298 gives permissible tensile and shear stresses for stainless steel and SCI, 1993 gives further guidance on the design of stainless steel fixings.  For more complex components in steel and aluminium guidance is given in BS 5950 and BS 8118 respectively.
 

Attachment to the support wall or frame
Concrete and masonry walls
The design of construction fixings is described in Section 04.09.
 

Support frames
Attachment to metal support frames falls into two broad categories:

• Simple attachment points may be fabricated as part of the steel or aluminium fabricated frame.  These are normally straightforward brackets or plates and are amenable to stress calculations for strength provided there are no unknown induced loads or movements.
• More complex attachments may be encountered as parts of proprietary systems.  In these, components are punched, pressed and folded to complex shapes and stress calculation is not possible.  However, these components are usually tested at the time the proprietary system is developed.

Whilst the strength of attachment points can be considered in this isolated way it may still be necessary to test them when integrated into the whole construction.  This is essential when the movement of the stone and any forces induced in the fixings are unknown.  It is also required if deflection of the attachment points is likely to affect the performance of the fixings or stone panels.

Testing of attachments to metal frames as part of the complete system is relatively straightforward.  The proof load for this component will be based on a lower factor of safety than most other parts of the system. The attachment points will normally be one of the first components to reach their proof load in any test.
 

Attachment to the stone panel
The design of the attachment to the stone panels will follow the same principles as attachments to a concrete or masonry wall. Failure of the attachment to the stone panel is manifest by fracture of the stone.  Accordingly tests are conducted at proof loads that incorporate a factor of safety of 3 or 4, and sometimes higher.  Some forms of fixing are able to carry this high load, for instance expansion or undercut anchors.  Other forms of fixing such as brackets or ties located with dowels may fail before the proof load on the stone is attained or may have so deformed as to have changed the load path and the loading on the stone panel.  For this reason it may be appropriate to fracture the stone using a stiffer and stronger connection.

Failure of the attachment results from fracture of the stone and the results from a series of tests may show considerable variation.  Some dense uniform stones will give consistent failure strengths whilst other stones will be highly variable in strength and fracture at vastly different attachment loads.  The greatest variability will occur for non-uniform stones, particularly those containing inclusions.

The strength of the attachment points will often depend on the dimensions of the stone panel tested as described below.  Test specimens will be subject to induced deviations but an attempt should be made to select specimens that are dimensioned so as to exhibit the least strength.  Dimensions of all test specimens should be checked and recorded.

Strength of the attachment may be affected by displacement and rotation of the fixings and by rotation of the panel as a result of flexure or dishing.  If this is the case then it is not normally possible to test the attachment to the stone panel in isolation.  Tests of at least a whole panel complete with all attachment details have to be conducted in this case.  For building stones greater than 20mm thick it is normal practice to ignore deflection of the stone panel and conducted tests on individual fixing points.
 

Kerfs
When testing the strength of a kerf detail it is important to use specimens with ‘shoulders’ of minimum thickness expected in the finished cladding.  This will give the lowest strengths under test.  The depth of the kerf and more importantly the position of the kerf plate or cramp in the kerf will also affect the strength under test.  The minimum anticipated penetration into the kerf should be used when setting up the test specimen.

If the kerf plate or cramp is sufficiently strong then the kerf may be tested using the fixing components to be used on the finished construction.  If the kerf plate or cramp is going to fail or deflect excessively before the proof load on the stone is reached then it is necessary to substitute a stronger fixing solely for the purpose of testing.  This should be of similar geometry to the original at its interface with the stone to create the same loads and load paths.

Where a single kerf plate secures two adjacent panels it is usually necessary to test in this configuration to avoid rotation of the kerf plate under asymmetric loading.

Wind loading on stone cladding may be both positive and negative.  If there is a significant difference in magnitude between the two loadings, or the kerf detail is strongly asymmetrical then the kerf should be tested in the weakest direction and for the greatest loading.  It may be necessary to test for both positive and negative load.
 

Dowels and wires
Dowels and wire ties lead to similar failures as those in kerfs.  The testing regime should be considered in the same way as for testing of kerfs, with one exception. Dowels and wire ties in particular are more flexible and their deflection under load may change the load paths and the loading regime in the stone.  It is difficult to substitute a stronger or stiffer component and still ensure the correct load transfer as for kerfs and tests have to be conducted using the true components.  Any unexpected or premature failure should be investigated as deflection of the dowel or tie at exceptionally high load may have invalidated the test.
 

Face anchors
Where expansion anchors, resin anchors or undercut anchors are used to connect to the stone panels they may be tested in accordance with BS 5080 Parts 1 and 2 for pull out strength.  However the tests will be into natural stone and the number of tests must be decided accordingly.

It should be remembered that with thin stone panels failure may also occur as a result of punching through of the anchor.  If the fixing detail requires the anchor to transmit compressive as well as tensile forces, the anchor should be tested under both positive and negative load.

Some undercut anchors are used with very thin stone panels and other thin materials.  In these uses they may be subject to bending as well as tension or compression.  In this case it is necessary to test the panel and all its fixings as a whole.
 

Through bolts
The testing of through bolts requires the same considerations as the testing of face anchors.  They behave in the same way and differ only in their method of construction.
 

Frame rebates
When a frame rebate is used to attach a thin stone panel it invariably induces stresses in the panel.  These details have to be considered with great care and should always be the subject of tests.  Testing may only be conducted on specimens comprising the whole stone panel and its fixing system.  Tests should include thermal and wetting tests to reproduce any dishing that may occur in the panel.
 

Fixings
Support angles
It is seldom necessary to test the load capacity and deflection of a support angle as these can normally be calculated.  If it is necessary to test then it will normally be easier, and more informative, to apply the loads with loading jacks.  It is of course important to apply the loads at the greatest anticipated offset from the support wall or frame.  If the loads are applied by loading jacks then a proof load can be applied to determine the factor of safety.
 

Support brackets
Brackets used to support the stone panels at one edge may be tested to determine their strength and deflection under load.  It may be necessary to test to ensure that excessive deflections do not occur and that the bracket does not contact the stone panel in some unintended way.

If the only concern about the brackets is one of strength then it will be possible to conduct the tests by applying loads to the brackets through loading jacks.

The deflection of any brackets will depend on their fixity to the supporting frame or structural wall.  If brackets are attached to a supporting metal frame or support rails then it will normally be necessary to test a complete frame with all brackets and loads present to determine the deflections of the brackets.
 

Ties and cramps
The tensile strength of a straight tie can be calculated but it may be necessary to test to check for prying of the tie from the back wall or buckling of the tie.  Ties and cramps can normally be tested as part of a complete assembly (panel, tie, anchor, kerf, dowel etc.) as they should reach their proof load before the stone panel or anchors fail.

When ties and cramps are tested alone any premature failure or excessive deflection should be investigated.  It may be that the observed failure mode is not the one that occurs when the stone panels and support frame or wall are present.
 

Face brackets
Face brackets attached to the inner face of the stone panel will want to rotate with the panel as it dishes as a result of temperature and moisture change and flexes as a result of wind loading.  For panels thicker than 20mm this is often judged not to have a significant effect on the panels and fixings.  For thinner panels it is necessary to test a stone panel complete with all brackets and fixings, and to investigate the effects of thermal and moisture change along with wind load, dead load and other applied loads.

For thin panels it is advisable to further test the assembly by subjecting it to cyclic wind loading, or an equivalent load such as 10000 cycles of full wind load.
 

Support rails
For support rails attached to brick or block masonry back walls the principal concern is local deformation of the support rails that may cause excessive deformation of the fixing brackets or ties.  Support rails may be tested separately from the stone panels by applying dead and wind loads through loading jacks attached to the support points, brackets or ties.
 

Stone panels
Stone panels may be tested for flexure but it is more normal to carry out bending tests on specimens of stone and calculate the bending strength of the panel.  This avoids the cost of preparing a statistically representative number of panels for test.  Test specimens of standard dimensions may be used but it is sometimes preferable to use test specimens of the same thickness as the proposed cladding. Samples should also have the same finish, particularly where the finishing process may affect the strength of the material near the surface.  A flexure test on a complete panel and fixings is difficult to perform, as the fixings will frequently fail before the proof load on the panel is reached.

Individual stone panels complete with fixings may be tested to show that the fixings are sufficiently flexible to not induce stresses in the stone and that the fixings do not deflect so far that they contact the stone panel in some unanticipated way.  When testing panels and fixings in this way it is important that fixings are not loaded asymmetrically if that is inappropriate and it may be necessary to test a larger specimen that contains several stone panels.  For instance panels fixed by rear face brackets may always be tested individually whilst panels attached by dowels cannot normally be tested separately.

Individual stone panels may be tested to determine what movement occurs when the temperature or moisture changes.  Panels complete with fixings may be subjected to a thermal cycling test, CWCT 1996, or a test that alternately wets and dries the stone.  Advice on the testing of cladding materials subject to thermal and moisture change is given in BBA :MOAT 22.  It may be necessary to test a larger specimen containing more than one stone panel if asymmetric loading on shared fixings is likely to affect the result.  Panels that are attached by back face brackets or a rebated frame and do not have fixings common to adjacent panels can always be tested as individual panels.

Panels may have to be tested using positive or negative wind loads or both.  Negative wind loads are generally applied using an air bag inserted behind the stone panel.  A method of test using an air bag is given in ASTM C1201 : 1991.  The loading regimes given in CWCT 1996 for curtain walling may be applied to such an air bag test.  At least one laboratory is able to apply negative wind loads using suction cups to pull on the stone panel.  These have to be articulated so that they rotate with the panel and do not induce stresses in it.  Positive loads may be applied by surrounding the panel and back wall with polythene or similar sheet and lowering the pressure in the cavity.  Alternatively it may be possible to apply the loads by laying the panel and support system down, outer face up, and using sand-bags to apply the required loadings.  The loading regimes to be used are the same as for negative loading.
 

Testing of large specimens
Large specimens should ideally represent the most critical conditions of the wall as actually constructed.  This is unlikely to be the case, and load tests are generally restricted to testing of particular aspects of individual components.

Large facade specimens may be constructed to test for air and water tightness in accordance with ‘Standard and guide to good practice for curtain walling’, CWCT 1996.  These specimens, which will be two or more storeys in height and typically one structural bay wide, may be used for the testing of natural stone fixings.  However it is unlikely that the proof loads on the stone panels will be achieved before some other aspect of the fixing or wall fails.  It is also unlikely that the test chamber used to load the specimen will be capable of withstanding the design wind load multiplied by a factor of four.

Large specimens offer the opportunity to load a single stone panel, or several stone panels within the specimen, using an air bag.  The adjacent panels will provide the correct edge restraint to the panel and fixings under test.  The disadvantage of this form of testing is that a single result is obtained to show the performance of a sometimes highly variable material.

Large specimens may be subjected to the cyclic temperature test described in CWCT, 1996.  This is the best way to check the effects of thermal movement as all panels in the specimen are heated and cooled.  The test will only show the amount of movement and dishing of panels that occurs.  It is not a satisfactory test of integrity as the induced stresses will be the actual stresses and no allowance will have been made for the load factor to be applied to stresses in the stone.  With a safety factor of four on the stone stresses, failure is not expected to occur at the operating temperature.  Moisture movement may also be assessed using a large specimen but again the induced stresses will be actual stresses and not factored stresses.

The one case for which testing of large specimens is useful if not essential is when the stone is used as a load bearing element that stiffens a support frame through plate action of the stone panels.  For constructions of this type the performance of the stone is critical and a comprehensive stone testing programme should be completed.  Full scale tests of complete frame assemblies and stone panels can then be undertaken to ascertain the failure modes and overall factor of safety.  An example of the structural use of stone is the use of prestressed Ketton stone for the Queen’s Building at Emmanuel College Cambridge, Structural Engineer, 1999.