Subscribe to our newsletter to receive the latest news and events from TWI:

Subscribe >
Skip to content

Mechanical testing - Tensile testing Part 2


Welding procedure approval for tensile testing

To approve a butt welding procedure most specifications such as ISO 15614 and ASME IX require tensile tests to be carried out.

These are generally cross joint (CJ) tensile tests of square or rectangular cross section that, as the name suggests, are oriented across the weld so that both parent metals, both heat affected zones (HAZs) and the weld metal itself are tested ( Fig.1). The excess weld metal in the cap of the weld may be left in-situ or machined off.

Fig.1. Square or rectangular cross joint tensile test piece
Fig.1. Square or rectangular cross joint tensile test piece

While it is possible to measure the yield strength, the elongation and the reduction of area of CJ specimens the fact that there are at least three different areas with dissimilar mechanical properties makes such measurements inaccurate and unreliable, although this is sometimes carried out purely for information purposes.

The specifications mentioned above require the UTS and the position of the fracture only to be recorded. The cross joint strength is usually required to exceed the minimum specified UTS of the parent metal. In most situations the weld metal is stronger than the parent metal - it is overmatched - so that failure occurs in the parent metal or the HAZ at a stress above the specified minimum.

In cases where the weld and/or the HAZs are weaker than the parent metal - welded age-hardened or cold worked aluminium alloys are a good example - this is covered in most specifications.

The designer must also take this into account in design calculations and provide some method of compensating for this loss of strength.

The tensile testing of flat plate butt welds presents few problems of specimen shape but those machined from a pipe butt joint are not flat and this curvature can affect the results. In the context of welding procedure approval testing, this is not significant since the test is used only for the determination of the UTS and the position of the fracture. For more accurate results the test piece may be waisted and may be machined flat as illustrated in Fig.2.

Fig.2. Flat cross joint tensile specimen machined from tube
Fig.2. Flat cross joint tensile specimen machined from tube

It may be necessary to machine a number of specimens through the thickness of a weld, particularly on very thick joints where the capacity of the tensile machine is insufficient to pull a full thickness specimen, Fig.3.

Fig.3. Multiple cross joint specimens machined from thick plate
Fig.3. Multiple cross joint specimens machined from thick plate

To test a small diameter tube, a solid bar is inserted in the bore of the tube to prevent the tube collapsing when the sample is clamped into the tensile machine.

Most weld testing is carried out with CJ specimens but longitudinally oriented specimens are useful particularly where the weld metal or the HAZ is very strong but ductility is low.

In a CJ specimen the parent metal can yield and finally fail without the weld metal or the HAZ experiencing any significant amount of deformation whereas in a longitudinal test piece the load is shared more equally.

A brittle weld or HAZ will not elongate with the parent metal but will crack, with the cracks opening, but not necessarily propagating into the parent metal, as testing proceeds.

The testing described above is that required by the welding procedure approval specifications. These provide no assurance that the welds in a structure will be suitable for their purpose such as elevated or cryogenic service and many application standards such as PD 5500 Unfired Pressure Vessels, and ASME VIII Pressure Vessels, require additional tests.

Since the strength of a metal falls as the temperature rises these specifications require elevated temperature tensile tests to be carried out at the maximum design temperature.

These tests are required to be carried out on the weld metal only and use a longitudinally orientated round cross section specimen from which an accurate measurement of the proof strength can be obtained.

Many application standards such as PD 5500 require tests additional to those required by, for example, ISO 15614-1. This must be remembered when procedure approval documentation is submitted for approval by the inspecting authority or the client.

Validity of tensile data

The samples taken are assumed to be representative of the bulk of the material but this is not always the case.

Tensile strength of a casting, for instance, is often determined from a specimen machined from a riser and this will have a grain size different from that of the bulk of the casting.

A rolled steel plate will be found to have different properties in the longitudinal, transverse and through thickness directions. Material specifications such as BS EN 10028, Flat Products in Steel for Pressure Purposes, therefore, require the tensile test to be taken transverse to the rolling direction so that the steel is tested across the 'grain' - the lower strength, lower ductility direction.

The size of a product can also influence the properties as, during heat treatment, the section thickness will affect the cooling rate with slower cooling rates, and hence softer structures, at the centre of thicker sections. This is dealt with in material standards by specifying what is known as the 'limiting ruling section', the maximum diameter of bar at which the required mechanical properties can be achieved at the centre.

In addition to variations of the properties due to the shape of the specimens and the testing temperature, the rate of loading will also affect the results.

Figure 4 shows how the tensile strength increases but ductility decreases as the testing speed is increased. The speed of the cross head of the tensile machine therefore needs to be controlled and BS EN 10002 specifies a stress rate range of 6MPa per second to 60MPa per second. The ASTM specifications have similar - but not identical - requirements.

Fig.4. Effect of speed of testing on strength and ductility
Fig.4. Effect of speed of testing on strength and ductility

Needless to say, calibration of testing equipment to guarantee operation within acceptable parameters is mandatory.

Relevant specifications

BS EN 10002 Methods of tensile testing of metallic materials.
BS EN 876 Destructive tests on welds in metallic materials - longitudinal tensile test.
BS EN 895 Destructive tests on welds in metallic materials - transverse tensile test.
BS EN ISO 7500-1 Tension/compression testing machines. verification and calibration of the force measuring system.
ASTM A370 Mechanical testing of steel products.
ASTM E8 Tension testing of metallic materials.
ASTM B557 Tension testing wrought and cast aluminium and magnesium alloy products.

This article was prepared by Gene Mathers.

This Job Knowledge article was originally published in Connect, May/June 2004. It has been updated so the web page no longer reflects exactly the printed version.

For more information please email: