ISO 9018 Tensile Testing of Cruciform and Lapped Welded Joints

What the test does

Cruciform or lapped test pieces are welded using the procedure being qualified, the fillet weld throats achieved are measured, and specimens of the required width are cut. The specimen is gripped so the load path crosses the transverse fillet welds and pulled to failure at 23 °C ± 5 °C unless something else is specified. ISO 9018 sets no rate of its own: it says the general principles of ISO 6892-1 apply, so the figure is read from there. Because only tensile strength and fracture location are wanted, ISO 6892-1 permits a single strain rate for the whole test, not above 0,008 s⁻¹ — on a 200 mm free length between the grips, up to 96 mm/min, its recommended 0,4 min⁻¹ working out at 80 mm/min. Where the elastic range is controlled by stress rate instead, the bands are 6 to 60 MPa/s for steel and 2 to 20 MPa/s below 150 000 MPa modulus: on a cruciform specimen 40 mm wide in 10 mm plate, 400 mm² of loaded section, 2,4 to 24 kN/s and 0,8 to 8 kN/s. Recalculate for your own section and free length. Two things are reported: the tensile strength, calculated on the area appropriate to the joint geometry, and the location of the fracture — through the weld throat, along the fusion boundary, or through the plate.

What it measures, and why it matters

Whether transversely loaded fillet welds carry the load, and where they give way. The interpretive difficulty, and the reason this method deserves its own standard, is that a fillet weld's load-carrying section cannot be seen once the joint is made. A full-penetration butt weld transfers load through the full plate section, which is measurable at any time. A fillet weld transfers it across a throat — the altitude of the weld triangle rather than a leg length — and a weld that looks generous can have a modest throat if it is concave. The strength depends on what was achieved, not on what was drawn.

Fillet welds under transverse load

A butt weld carries load through its full section. A fillet weld carries it through a throat that nobody can see once the joint is made.

Joint types

Cruciform and lapped

Welds

Transverse stressed fillet weldsThe load crosses the weld rather than running along it.

Applies to

Plates, sheets, extruded bars and other solid sectionsThe standard uses plate as a general term for all of these.

Reported

Tensile strength and fracture location

Evaluation of results

Explicitly not included in the standardNo acceptance criteria. Those come from the application standard or the contract.

Measure the actual throat, not the specified one

DakFillet weld throats vary with fit-up and technique, and the strength depends on what was achieved rather than what was drawn.

Fracture location matters as much as in a butt joint. Failure through the throat is a weld result; failure through the plate means the welds were stronger than the material they joined.

Test speed

Where the rate comes from

ISO 6892-1. ISO 9018 sets none of its ownClause 4: unless otherwise stated, the general principles of ISO 6892-1 apply. Test temperature 23 °C ± 5 °C unless otherwise specified.

One strain rate throughout

Not above 0,008 s⁻¹, since only tensile strength is wantedOn a 200 mm free length between the grips that is up to 96 mm/min; the recommended 0,4 min⁻¹ works out at 80 mm/min. Recalculate for your own free length.

Or by stress rate in the elastic range

6 to 60 MPa/s for steel; 2 to 20 MPa/s below 150 000 MPa modulusOn a cruciform specimen 40 mm wide in 10 mm plate — 400 mm² of loaded section — that is 2,4 to 24 kN/s and 0,8 to 8 kN/s. Our arithmetic on a stated section, not a figure from either standard.

Reported

Tensile strength and fracture location

Acceptance

From the application standard, not from here

Section a failed joint to see the root

DakLack of root penetration in a fillet weld is invisible from outside and is a common cause of low throat failures.

What the machine must be capable of

Substantial force — a cruciform coupon in structural plate commonly needs a hundred kilonewtons or more — with wedge grips sized for the plate thickness and enough daylight for the assembled joint. Alignment good enough that the load is genuinely axial matters here more than on a plain coupon, because the geometry already carries an offset and a bending component added by the fixture will move the failure to whichever weld happens to be on the tension side.

Grips and fixtures for this method

TJ-144

Heavy Duty Hydraulic Grips

Heavy duty hydraulic wedge grips hold a constant clamping force on a thick welded coupon as it necks, which a mechanical wedge does not always manage at these loads.

Specifications

Self-tighteningTJ-15

Universal Parallel Wedge Grips

Universal parallel wedge grips with flat inserts, where the plate thickness and failure load sit within their range.

Specifications

How the test runs

1. 01Prepare cruciform or lapped test pieces to the specified geometry.
2. 02Weld them using the procedure being qualified, recording all parameters.
3. 03Measure the actual fillet weld throats achieved.
4. 04Cut specimens to the required width.
5. 05Fit wedge grips sized for the plate thickness.
6. 06Align the specimen so the load path is as intended for the joint type.
7. 07Load at the specified rate to failure.
8. 08Record the maximum force.
9. 09Identify the fracture location — through the throat, the fusion boundary, or the plate.
10. 10Section a failed joint where root penetration is in question.
11. 11Compare against the acceptance criteria in the application standard, naming it.

Calculations

What the report has to contain

• Reference to ISO 9018 and the edition
• Parent material and welding procedure identification
• Joint type — cruciform or lapped
• Test piece geometry and specimen width
• Specified and measured fillet weld throats
• Rate of testing
• Maximum force and the calculated strength, with the area basis stated
• Fracture location, classified
• Any evidence of lack of root penetration
• The acceptance criteria applied and the standard they came from

What goes wrong in practice

Reporting a pass or fail as though the standard supplied one, when it explicitly does not — the acceptance criteria come from the application standard or the contract and their source belongs in the report. Reporting a strength without the fracture location, which discards the diagnostic half of the result. Assuming the specified throat is the throat achieved. And failing to section a low-failing joint, which is usually where an unfused root — invisible from outside — turns out to explain everything.