Start with the weld question

Phased array ultrasonic testing (PAUT) uses electronically controlled groups of transducer elements to steer and focus sound. Multiple beam angles can examine an agreed weld volume and produce views that help locate reflectors. It is not a photograph of every flaw, and a colourful scan is not itself an acceptance decision.

NDT Australia's ultrasonic testing service includes conventional and phased array techniques where suitable. Before choosing PAUT, provide the joint drawing, material, thickness, weld process and examination objective. A fabrication acceptance examination and an investigation of an in-service indication may need different procedures and reporting detail.

Geometry determines the scan plan

The inspector needs to consider bevels, root configuration, cap profile, adjacent attachments and probe travel on each available surface. A scan plan should show which beams interrogate the target region and where geometry creates blind or weakly covered areas. Reflector orientation matters: a discontinuity that does not return sound towards the probe can be difficult to detect.

Single-sided access may be workable, but it must not be described as equivalent to unrestricted access without demonstrating the required coverage. Thin sections, coarse-grained materials, curved surfaces and complex connections can require different probes or complementary methods. Agree inaccessible zones before attendance rather than discovering them at handover.

Preparation and calibration are part of the result

Provide clean scanning surfaces with enough room for the probe and any encoder. Couplant, surface roughness and temperature affect transmission and repeatability. Coatings require consideration under the procedure; do not assume they can remain simply because the equipment produces a signal.

Calibration and sensitivity checks must suit the material, thickness range and technique. Relevant reference reflectors, equipment settings and checks before and after examination give the data a traceable basis. An encoded scan can retain positional evidence, but position tracking and coupling still need control; data storage does not compensate for a poor examination.

Separate indications from acceptance

Record the weld identifier, examination extent, indication position and any sizing basis. Geometry echoes and material responses may resemble discontinuities and need interpretation. PAUT sizing has uncertainty; a reported indication length or height should be read alongside the technique and its limitations, not as an exact physical dimension.

Acceptance comes from the agreed specification and relevant procedure, not from the display alone. An engineer may need to review the significance of an in-service finding against loading, history and design information. Surface testing or another ultrasonic approach may add evidence where the original scan leaves a question unresolved.

Ask for a reviewable data package

For a fabrication weld, a useful package might combine a weld map, scan plan, calibration record, located indications and retained scan files when commissioned. This is an example of the evidence structure, not a promise that every job will use encoded acquisition or identical outputs.

Agree file formats, interpretation responsibilities and how uncovered regions will be reported. If a repair is undertaken, preserve the original record and identify the repaired area in the repeat examination. Comparing scans is meaningful only when location, preparation and technique differences remain visible.