Testing Of Stone (show list of test data)

Stone has been used as Building Material for thousands of years. Originally  stone structures were solid. Veneer cladding of concrete or brick structures developed after the early 1900s.

Since the 1960's the thickness of stone veneer on buildings has decreased. Today a common thickness is 30mm with some cladding as thin as 20mm.

This trend towards the use of thinner stone was not, as one would expect, a  consequence of increased knowledge of stone material properties.

It was driven largely by economic factors. As a consequence several major  projects worldwide suffered substantial problems from using thin veneer cladding that had not been adequately researched. This in turn has driven the research, testing and setting of standards that now predominates the stone industry.

When thin stone panels (i.e. less than 50mm) are used without full knowledge and control, problems can occur with the cladding (cracks, spalls, fractures,  bowing, collapse). Most commonly these types of problems occur in stone panels at connections, reinforcing the need to address connection design as well as flexural design.

The above does not imply all stone usage requires testing. Table 1 details possible testing formats for various situations.

Any testing program should be considered in 4 phases with each phase being of  varying importance depending on the circumstances.

Selection Testing
The selection of possible materials will generally be based on aesthetics  however, the structural suitability of the possible materials also needs to be assessed.

This could be based on historical data for smaller projects or a full blown  geological examination of possible deposits with associated testing of derived samples for large scale facades. This latter process may require the services of  a stone consultant.

Initial Testing
Prior to a full design being carried out it is necessary to have data on which to base such a design. This data could be that provided during a selection testing program if such is sufficient or it may require more exhaustive testing on the preferred material hence a second phase or initial testing phase.

Both of the above processes use laboratory bench tests to assess the relative  properties of samples (tests such as compressive strength, flexural strength,  absorption, etc)

Design Testing
Having carried out the preliminary design and arrived at panels of suitable  thickness and size, it is then necessary to fully test the overall design and details with testing such as:

   1. Kerf or Fixing Detail Testing - Structural Connection Tests
   2. Full size Structural Test - Panel Testing
   3. Service ability Tests - Air Infiltration, Etc

Construction Phase Testing
Having established the parameters for use of the material it is then necessary to ensure that the material properties remain consistent and exceed  the values set for acceptability. Frequency and types of testing need to be set down as do procedures for re-testing of failures, quarantine methods for unacceptable material etc.


DEALING WITH EACH OF THE ABOVE PHASES IN GREATER DETAIL

SELECTION TESTING
Suitable resources need to be identified for a project both in terms of  quality and quantity. It is not unusual to appoint a stone consultant or geologist to examine such options or confirm that the aesthetic choices are satisfactory for use.

Part of this process should be to assess the structural capabilities of the  material/s.

This testing does not need to be exhaustive as it is aimed at eliminating  structurally risky materials from further consideration rather than fully  proving suitable material.

It is suggested that testing be limited to Flexural Strength plus Absorption and Specific Gravity at this stage.

As an indication on whether a material should be considered further (or not) reference should be made to the "The Minimum Requirements for Dimensions Stone" as per ASTM C615 (refer Table 2). Obviously comparisons can be made between different materials under consideration i.e.; some higher strength materials may not require the same thickness to be satisfactory hence promoting possible savings.

INITIAL TESTING
It will be necessary in most circumstances to have reliable structural information relating to the stone on which to base any design.

Historical data on the stone may give some indication on strength and  consistency of results but information relating to currently available material is necessary in most instances.

Selection of testing samples needs to be addressed at this stage especially  if the possible materials are still at a preliminary stage. Again on major projects the advantage of a specialist consultant can be seen. As a minimum the samples should be drawn from normal production block/s extracted by using the quarry technique to be employed for the project.

Directionality of the material also needs to be considered prior to any  ongoing testing program being in place. Stone is non-isotropic; its properties vary depending on which direction is being considered. This is as a consequence of the method of deposition. In sedimentary rocks, (sandstone etc) layers or beds are sedimented horizontally or bedded. Geological activity may alter these layers from horizontal, however a difference will continue to exist in the properties along the bed compared to across the bed.

This is the same with Metaphoric Rocks (Marbles) which are altered sedimentary structures (heat and pressure). Igneous rocks (Granites) have a similar phenomenon caused during the cooling of the structure. Lateral stresses are set up as cooling proceeds. Micro cracking predominant in a certain  direction or directions will result from stress relief. This will affect the properties in certain directions.

Most experienced quarry masters will tell you that deposits will split more  easily in a certain direction. This is the Rift of the stone and is a consequence of the directionality referred to above.

This directionality may be significant or may be of little consequence however an assessment needs to be made on its possible effects, hence initial testing needs to take this into account.

At this juncture the visual appearance of the stone in various directions  needs to be assessed and decisions made on which orientation, if any, is  preferred. This visual effect is directly related to the directionality of the  properties and also can be highly visible or of no consequence.

Obviously no hard and fast rules regarding visual orientation can be put in  place at this time as it may conceivably be that the most visually satisfying orientation could be the weakest orientation and hence could become the most expensive. However it is another matter to consider during initial testing.

As indicated previously the testing carried out is generally in accordance with the following American Standard Test Methods (ASTM's).

FLEXURAL STRENGTH (ASTM C880)
Flexural strength is a key property needed for structural design of dimension  stone. ASTM C880 flexural strength testing is performed by applying loads at the quarter points of a stone specimen cut from a larger panel or slab. The load is increased gradually until the test specimen fractures. The maximum applied load is recorded and the flexural stress that occurred in the specimen at fracture is calculated. Five samples are tested for each of the four conditions of testing  (wet or dry and parallel or perpendicular to fit).

ASTM C880 allows for testing stone of various thickness. It requires that the  span length for testing be 10 times the thickness. It currently requires that  the specimen face in tension have a fine abrasive finish, but it does allow for reporting of variations from the specified procedures. Hence, stone cut to production thickness and with anticipated production finish, can be tested. An upcoming edition of C880 will more explicitly specify testing of stone at production thickness and finish and increases the width of the specimen to 100mm."

Obviously it makes sense once the critical factors are known to concentrate  testing on these. i.e., If wet testing results in lower results than dry  testing, then do not carry out dry testing. Similarly once the critical direction of testing is known, this should be concentrated on. This however depends on visual orientation and directional properties and care should be  taken not to drop multi-directional testing too early. It should be noted that to fully understand the micro-structure of the material testing in 3 orientations may be necessary initially.

MODULUS OF RUPTURE (ASTM C99)
"A measure of the flexural strength of stone specimens can also be obtained  from ASTM C99 Modulus of Rupture test. Specimens of stone for this test are specified by ASTM to 100mm wide x 200mm long x 50mm thick, with a fine abrasive finish. The test load is applied at a single point in the centre of a 180mm span. ASTM C99 specifies testing of five specimens wet and dry and with a load parallel and perpendicular to the rift plane. Although both the ASTM C880 Flexural Strength test and the ASTM C99 Modulus of Rupture test provide a measure of bending strength, they do differ. The ASTM C99 test specimen does not  take into account the actual thickness or exterior finish of the stone to be  used on the building and has a shear component which stiffens that sample, giving a higher apparent strength than the actual bending strength. ASTM C99  (Modulus of Rupture) should not be used for design purposes. In addition, the  centre point loading configuration in ASTM C99 typically causes the failure to  occur directly under the applied load. ASTM C880 specifies that the test load be applied at the two quarter points of the specified and allows for failure to occur at its weakest point between the load points. ASTM C880 provides results  that compare more closely with the result of full panel flexural tests that does  ASTM C99.

COMPRESSIVE STRENGTH (ASTM C170)
Compressive strength of stone specimens is determined in accordance with ASTM  C170. Compressive strength tests are performed on saw cut cubes or on core-drilled cylinders. Tests are performed by loading the specimen in a  calibrated test machine until it fractures. The maximum applied load is divided by the loaded area of the specimen to calculate the compressive strength of the specimen tested". Usually five samples in the wet and also dry conditions are used.

"Although the structural design of a thin stone panel is not normally  controlled by compressive strength it is still important to know that the compressive strength of the stone under consideration meets the minimum  properties specified by ASTM and to know if the properties of the stone are consistent with historical values. A designer should seriously consider his  actions before specifying a stone that is known not to meet the minimum strength requirements specified by ASTM."

ABSORPTION AND SPECIFIC GRAVITY (ASTM C97)
Absorption and bulk specific gravity of stone specimens are determined in  accordance with ASTM C97. Bulk specific gravity is also referred to as density. Specimens are first dried by heating and then weighed. They are then immersed in water for 48 hours to become saturated, they are weighed while submerged in  water and then weighted in air while saturated. By comparing the dry buoyant and saturated weights obtained from these three weighings, the amount of water absorbed and the specific gravity of the specimens is calculated. This test is performed on three specimens for each stone under consideration. These test specimens can be the same size as those used for compressive strength testing. In fact, if material is limited, the specimens used for absorption and specific gravity tests can later be used for compressive strength tests.

As with Compressive Strength, absorption and specific gravity have little  direct relationship to the structural design of a dimension stone on a building. However, these tests should be performed to verify that the stone meets ASTM specified physical requirements. It should be noted that experience has shown  that higher absorption values may mean poor weather ability of the stone.

Other standard tests that may be of significance are:
 

1. Determination of Co-efficient of Thermal Expansion in accordance with either ASTM E381 or line with the Australian Standard (to be issued). 2. Abrasion Resistance - ASTM C241 - (This test is currently under review but measures resistance to abrasion from foot traffic (i.e. resistance to wear). 3. Slip Resistance - in accordance with AS/N\Z 366.1-1993. This assesses the slip resistance of stone under wet and dry conditions setting limits on acceptability of materials in relation to slip resistance. 4. Polished Friction Value Materials for Pedestrian Traffic. This test is a measure of the slip resistance of material before and after extensive use (i.e.  polishing effect by pedestrian traffic). The test is carried out as per AS1141.41 modified in accordance with Council of the City of South Sydney requirements.

QUANTITY OF TESTING
Some of the above tests may only need to be conducted once (i.e. Compressive Strength, Co-efficient of Thermal Expansion, etc) and subject to a satisfactory  result further testing would not be required.

Other tests (predominantly Flexural Strength) require a reasonable initial base to assess consistency and predict acceptable safe limits. Here the question  is the amount of initial testing, which is a statistical one and will vary  depending on the material. A rule of thumb may be 1 test specimen/100 panels to 500 panels.

DESIGN TESTING
It is assumed that wind tunnel test would have been carried out where  necessary to indicate maximum transverse load across elements. From here it is relatively simple to arrive at the notional tensile strength required of a  cladding panel for a given panel size and thickness and check this against flexural strength results achieved for this material. This will indicate a factor of safety which may be considered satisfactory of otherwise.

Similarly it is normal to carry out finite element analysis at point of  connection or using an empirical formula, work out the stress at the connection.  Such an empirical formula is the one relating to sawn kerfs:

i.e.: f = 38.6W 1 + 0.75 d 152h + 16.25

2r d² 1 (h x d)

Where f = bending stress capacity of connection (MPa)

W = load applied to kerf expressed in (KN per mm)

d = effective thickness of stone at kerf (mm)

r = inside radius of the corner of kerf (mm)

h = distance between applied load and point of maximum stress (mm)

This wold also be compared to the results from flexural strength test results  and the resultant Factor of Safety reviewed.

Whilst the analytical or empirical approach is satisfactory the actual stresses in a panel are complex, transient and affected by support  conditions.

For these reasons the following design tests are often carried out

> Full size panel testing
> Structural connection testing

FULL SIZE PANEL TESTING
A series of panels are tested in a test frame that simulates in service  conditions of support, load reversal, ambient conditions etc.

Care should be taken to duplicate the panels most likely to fail in service, i.e. thinner stone (within range to be expected), in highest wind load area and/or with worst stress concentration configuration.

Normally 5 (minimum) identically sized panels will be tested to ensure some  variation in properties and the full sized samples will be produced along with Flexural Strength samples which will also be treated to ensure comparability of results.

Normally transverse loads are applied and reversed across the panel to 1.5 times maximum design load and held at that level for a specified period. Subsequently the panel is loaded to failure and the load recorded and checked against Design Loading and Flexural Strength Results and anomalies commented on.

STRUCTURAL CONNECTION TESTING
The designated connection design or a series of possible designs are tested  to destruction to assess load characteristics and these are compared to  empirical or finite analysis results and any anomalies commented on.

Always remember to allow sufficient tolerance in connection design to account  for any panel, frame and location tolerances, i.e. kerfs should be deep enough to avoid bottoming out fixing rebates to panels should account for frame tolerances, etc. This is necessary to avoid refiguring fixing details later on and possibly having to retest.

In the past test results from connection tests and full sized panel tests  have shown excellent correlation to the empirical formula for kerfs and the Flexural Strength test results respectively, indicating these are good approximations for design testing and can be used to arrive at stone thickness and connection prior to design testing being used to verify results.

The services of a professional testing organisation should be enlisted to  program any design testing. Methodology should be exactly detailed and followed  to ensure repeatability and consistency of results. Natural Stone results will vary enough without adding further variables.

Other design testing that can be carried out are Serviceability Tests  including Air Infiltration and Water Infiltration tests. These are carried out on a full scale mock. As well Sealant Compatibility, Adhesion and Staining Tests to evaluate the most suitable sealant for the project should be done. These are of much broader application than just stonework and consequently are not covered in detail here.

CONSTRUCTION PHASE TESTING
It is vitally important that once the design phase is completed and supply has commenced that the material is kept under scrutiny to ensure the design is not compromised. It is obviously not necessary to conduct most of the tests  previously detailed on an ongoing basis, however it is recommended that the flexural tests be maintained during construction on a regular basis and also that absorption and bulk specific gravity be checked occasionally.

The frequency of testing depends on the circumstances involved such as material variability, size of the project, risk involved, etc, however the basic principle should be put to more effort into testing early to prove consistency  (or otherwise).

It is also important that testing be flexible to allow for changing  circumstances i.e., less testing if results consistently exceed hurdle by a  large margin or more testing if material starts to show inconsistency or lower  results.

Some provision should also exist to carry out more testing through the project when conditions change, i.e. extraction from different part of quarry.

Adjustments can be made also to test procedures to minimise testing effort whilst still achieving verification requirement.

i.e. Delete dry testing if not as critical as wet testing.

Delete perpendicular testing if parallel is critical (except where quarry  change occurs).

Alter number of specimens from 5 to 3/sample.

A TYPICAL CONSTRUCTION PHASE MAY BE:
First 10% Flex Test each block in 2 directions 5 specimens/sample plus absorption and bulk specific gravity.

Second 10% Flex test every 3rd block in 1 direction 3  specimens/sample.

Third 10% Flex test every 5th block in 1 direction 3 specimens/sample plus absorption and bulk specific gravity.

Fourth 10% Flex test every 10th block in one direction last 10% 3 specimens/sample.

Above equates to 22% of blocks tested but by virtue of high early testing  content will indicate consistency or otherwise of material early on.

INTERPRETATION OF TESTING
By one means or another a hurdle will have been set for acceptance or  rejection of the test results compiled during the construction phase. Normally Flexural Test Results are interpreted by the mean of the results or by  interpretation of the mean and the standard deviation for each sample.

i.e.: "A sample shall be considered acceptable if the mean result of that  sample equals or exceeds10MPa and the mean less 1 standard deviation of the  sample equals or exceeds 7.5MPa."

It should be noted also that some credence has given to single sample  results

i.e.: "A sample shall be considered acceptable if all sample results equal or  exceed 7 MPa."

This is not recommended as a highly variable material may still satisfy this criteria and the testing is not really not assessing the consistency of material.

PROBABILITY
Statistical analysis predicts that based on mean of the sample that 50% of the sample will have results lower than the acceptance level. For analysis based on Mean less 1 standard deviation 16% or results will be below acceptance level.

This can be carried further i.e.:

Acceptance level Volume of population expected to be excluded

Mean 50%

Mean - 1 standard deviation 16%

Mean - 2 standard deviations 2.25%

Mean - 3 standard deviations 0.15%

Analysis based on Mean - 3 standard deviations then encompasses 99.85% of a  sample population. As a consequence the factor of safety involved should be  substantially reduced over that involved with an analysis based on simple mean.

On large projects it is recommended that a probability analysis be carried out, based on Initial Test Results, to assess the hurdle level required for construction phase testing.

FAILURE & RETESTING
It will always be the case that some samples will not pass the hurdle set and  some format should be in place for retesting. It makes sense that at the time of  taking samples for construction testing further material is put aside for duplicate testing.

Tested samples should also be maintained for further inspection for the  period until results are approved as low results and retesting may not be necessary if an obvious fault is detected in a sample. By inference sample selection should reflect production methods for selection of panels.

SAFETY FACTORS
Once the wind tunnel tests have been conducted and the design pressures are  known, once the kerf design and bending design have indicated maximum stresses to be expected for nominal panel sizes, thicknesses and kerf details, then it is a case of deciding whether the material is suitable for use. This is based on whether the Safety factor is sufficient.

Many factors will contribute to Safety Factor assessment. Some of these  are:

Stone is non hetro geneous - it is not uniform and properties vary from  location to location.
What's more even this variability is variable from type to type.

Stone is non isotopic - it does not have equal properties in all directions.
Stone can and will deteriorate with age (with levels of deterioration  depending on stone type and conditions).
Stone is brittle and any cracking usually results in a complete failure.
Uncertainty of Loading (if any).
Size of test sample - if small, safety factors should increase.
It is recommended that a full analysis be carried out on Safety Factors based  on initial test results and other information (historical data, etc). As a basis to start assessment the historically accepted factors of safety have been:

Granite - 3

Anchorage Components in Granite - 4

Fixing Systems Generally - 4

Marble Veneers - 5

Limestone Veneers - 8

An example of possible analysis is attached.

SAFETY FACTOR ANALYSIS

TEST DATA

No of samples n = 15

Mean Strength x = 12.4MPa

Standard Deviation < = 1.67Mpa

Coefficient of Variability V = 13.5%

SAFETY FACTOR CALCULATION (ASSUME 30% STRENGTH LOSS DUE TO WEATHERING)

x - 3* = 12.4 - 3 x 1.67 = 7.39MPa

Variability Factor VF = x= 12.4 =1.68

x -3< 7.39

Weathering w = (x - 3<) (1.00 Minus  Loss Due to Weathering)

= 7.39 x 0.7 = 5.17MPa

Weathering Factor wF = (x - 3<) = 7.39 = 1.43

W 5.17

Wind Factor = WF = 1.17 (From Table 4)

Safety Factor = VF x wF x WF

= 1.68 x 1.43 x 1.17

= 2.81

ALLOWABLE DESIGN STRESS Fb = 12.4 = 4.41

2.81

This analysis can be rechecked statistically using the attached Table  5 and using a Wind Load Factor Maximum of 1.17

i.e.: Fb = x- N Where N is number of standard deviations for

specified confidence levels and risk factors
In our example Fb = 4.41

x = 12.4

* = 1.67

n = 15

This implies N = 4.33

From Table 5 this would indicate a confidence level of 50% and a risk factor between 1/10,000 and 1/100,000

This may not be an acceptable factor of safety. Based on say 90% confidence  and a risk factor equal to 1/1,00,000, a more acceptable result for 15 samples  would indicate N = 5.77.

This would indicate Fb = 12.4 - 5.77 x 1.67

1.17

= 2.36MPa

i.e. Safety Factor = 12.4 = 5.2

2.36

TABLE 1 POSSIBLE TESTING FORMATS