The engineer is often required to ‘certify’ that:
the steel materials procured for the project meet the performance intent of the relevant Australian material Standards called up in the project specification, and
the fabricated steelwork structure, in the form of the completed building or structure, meets the performance intent of the relevant Australian Standards, which usually would be AS/NZS 5131.
Neither of these responsibilities is trivial and both must be undertaken after the engineer fully informs themselves in order to satisfy duty of care under WHS Regulation. Properly ascertaining compliance of either materials or fabricated steelwork is a specialist undertaking that few engineers are specifically trained for and that contracts often do not properly address, either in terms of scope or allowance.
The sections below provide guidance on where and how to start the process of ascertaining compliance of steel and steelwork.
Our design Standards provide guidance on requirements for material compliance to suit the performance intent of the Standards:
Requires test reports or test certificates be provided by the manufacturer or by an independent laboratory accredited by signatories to the International Laboratory Accreditation Corporation (Mutual Recognition Arrangement) (ILAC MRA) or the Asia Pacific Laboratory Accreditation Cooperation (APLAC) on behalf of the manufacturer. Where there is a dispute as to the compliance of the steel, reference testing shall be carried out by independent laboratories accredited under ILAC MRA or APLAC.
The veracity of the testing laboratories must be established. There have been documented cases of laboratories undertaking testing they are not competent for.
"If unidentified steel is used ... unless a full test in accordance with AS 1391 is made, the yield strength of the steel used in design shall be taken as not exceeding 170 MPa, and the tensile strength used in design shall be taken as not exceeding 300 MPa."
If the steel cannot be properly identified according to the requirements of AS 4100, this very significant reduction in strength must be utilised in design. For most project scenarios, this strength reduction will be untenable as a design requirement, particularly if the design and sourcing of product has already been undertaken.
The marking of steels shall be as specified in the appropriate Standard.
Similar requirement to Clause 2.2.3 in AS 4100 above.
"Where steel does not satisfy the requirements of Clause 2.2.2, it shall be classed as unidentified steel. Unidentified steel shall not be used."
This clause is more onerous than the similar Clause 2.2.3 in AS 4100. No strength reduction is allowed, reflecting the importance of steel veracity to bridge performance.
The material Standards (AS/NZS 1163, AS/NZS 3678, AS/NZS 3679.1, AS/NZS 3679.1) provide consistent guidance across all four Standards, namely:
Clause 11.1 Identification: itemises the range of requirements for the identification marks to be provided on the steel product. Products not marked with the provisions are non-compliant with the Standard.
Clause 11.2 Test and inspection certificates: outlines the requirements for test and inspection certificates that shall be available to the purchaser for each batch produced. Tests shall be undertaken by laboratories accredited by signatories to ILAC MRA. Minimum sampling and testing procedures shall conform to Appendix B.
Appendix B Product Conformity: is a normative appendix that outlines the scope for initial type testing (ITT) and factory production control (FPC) to be demonstrated by the manufacturer or supplier. Minimum sampling and testing frequency plans are documented. The product conformity requirements are intended to enable conformity assessment to be made by a manufacturer or supplier (first party), a user or purchaser (second party) or an independent body (third party).
Requires that if documentation is inadequate or incomplete such that compliance cannot be adequately established, the steel material or component shall be treated as unidentified steel in terms of AS 4100, AS/NZS 4600 and AS/NZS 5100.6. If testing of nonconforming material or components is undertaken, the results shall be in accordance with Clause 13.3.8.
"The type of testing shall be consistent with the type of testing specified in the relevant product Standard cited in Clause 5.3, and shall be sufficient to establish a proper statistical basis. Single or limited test results shall not be acceptable. Appropriate statistical processes shall be applied to multiple samples in order to establish confidence in the properties of the product."
This clause makes it very clear that a proper regime of testing needs to be undertaken (refer to the following section).
Where testing of steel is required, the procedure for and extent of sampling must be considered and carefully managed to ensure the intent of the Standards is maintained. A proper statistical process should be applied consistent with the intent of the Standards.
Clause C2.1 of the Commentary to AS 4100 speaks to the yield stress and tensile strength used in design and states, in part:
"Variations in the yield stress and tensile strength, which occur during manufacture, are accounted for in the derivation of the capacity factor (Ø) (see Clause C3.4). Because of this fact, the actual values of yield stress or tensile strength recorded on mill test reports or certificates cannot be used for design. The values used in Table 2.1 should not be exceeded in design or else the derived capacity factors of Table 3.4 are rendered invalid.”
This makes the point that actual values of yield stress and tensile strength as shown on test certificates are not appropriate to be used for design purposes.
ASNZS 5131 requires that, with regard to testing of nonconforming steel or products, “appropriate statistical processes shall be applied to multiple samples in order to establish confidence in the properties of the product”. This clearly establishes that single or limited tests are not appropriate.
When undertaken as part of Factory Production Control (FPC) as part of the manufacturing process, tests of the steel are undertaken at regular intervals and build up a statistical profile of the manufacturing performance of the steel plant as well as the particular properties of the steel batches produced. For example, as would be expected, there is a variation (often a normal distribution but depends on the property measured) around a mean value of the yield stress for each grade of material produced, as indicated in the figure below:
The figure indicates an actual distribution of long-term yield strength test results from a quality Australian manufacturer for a nominal grade 300 MPa steel. Note the mean value is significantly higher, at approximately 355 MPa, compared to the nominal grade of 300 MPa, reflecting the fact that the grade designation is designed to meet the 5% characteristic value statistically.
Our structural steel design Standards, such as AS 4100, rely on guaranteed values of chemical composition, mechanical properties, dimensional tolerances and methods of manufacture in the nominated product Standards. The capacity factors nominated in Table 3.4 of AS 4100 were derived using statistical analysis of results from steel testing of material complying with the nominated product Standards using the normal distribution curves obtained from the manufacturers as at the date of preparation of AS 4100. The capacity factors in AS 4100 are therefore calibrated against normal distribution curves as shown in the above figure, using the 5% characteristic value as the assumed design value.
It therefore follows that if testing of steel is to be undertaken to establish design properties for use in design Standards, the testing regime must:
establish all properties required to be documented under the product conformity requirements of Appendix B of the relevant product Standards, AS/NZS 1163, AS/NZS 3678, AS/NZS 3679.1 or AS/NZS 3679.2 as applicable
have samples taken from and representative of the same mill and product designation (size) as reported on the material certificates that are supplied with the steel product. A separate sampling and testing regime must be undertaken for each collection of same product designation and same mill source. A statistical analysis of test results is meaningless if representing a range of different types of material, perhaps from different mills
have sufficient tests of any one property to allow a proper statistical basis to be established. AS 5104 ‘General principles on reliability for structures’ Appendix D provides guidance on appropriate statistical techniques for design based on testing.
Feedback in the market indicates that limited testing of materials that have not been produced to Australian Standards has been used to, in effect, alter the original design grade of the material. The practice typically involves taking a single or very limited number of tensile test results and using these as the sole basis for nominating a steel grade that is (usually) higher than the original intended nominal design grade.
As noted in the section on ‘Testing of Steel’, the use of a single or limited test results is not consistent with the intent of our steel product Standards or the theoretical basis for our design Standards. The testing regime described in the section on ‘Testing of Steel’ must be implemented, assessing a design value based on the 5% characteristic value.
A related issue is that of ‘grading by selection' (PDF) practiced by some less reputable mills and distributors in the supply chain.
Steel Innovations paper 2013 (PDF)