Automated UT (AUT) for pipeline girth welds

On a lay barge or a land spread, time is billed literally by the minute. Once the mainline girth weld is complete, how fast testing ends and the "accept" or "cut out and repair" call is made sets the rhythm of the whole spread. That is the real reason automated ultrasonic testing (AUT) has largely displaced radiography (RT) in modern pipeline construction. AUT delivers an almost immediate evaluation after a single scan pass — on the order of one to two minutes — with no radiation exclusion zone, a permanent digital record, and an output that feeds the accept/reject decision directly. This article covers how mechanized and automated UT work, the zone discrimination approach, the use of phased array probes, calibration and recording discipline, and the acceptance of pipeline girth welds, seen through the eyes of an NDT Level 3 and grounded in field reality.

This article is for educational purposes and does not replace the official standard, the procedure, or the judgment of qualified personnel. Always work from the current edition of the applicable standard and the project procedure.

1. Mechanized UT versus automated UT

The two terms get mixed up constantly in the field. In mechanized UT, the probe travels on a mechanical scanner and an encoder records its position continuously — but the inspector performs the evaluation. Automated UT adds a decision-support layer: software handles data acquisition, threshold comparison, and the alarm/flagging logic. In practice, the systems used on pipeline girth welds combine both. A probe band makes one revolution around the weld, the encoder stamps position every millimeter, and the system builds a separate channel of data for each weld zone.

The critical point is this: AUT is not a black box. A competent inspector makes the final acceptance decision; the system only presents the data in a repeatable, position-referenced and auditable form. Under ISO 9712, data acquisition and the conduct of the testing require Level 2 personnel as a minimum, while technical approval of the procedure, the zone and focal law setup, and the calibration block configuration is a Level 3 responsibility. This keeps responsibility for the application cleanly separate from responsibility for the technical design.

2. The logic of zone discrimination

Zone discrimination is the heart of pipeline AUT. It divides a narrow-gap weld preparation into horizontal slices stacked through the pipe wall — the zones: root, hot pass, fill zones (fill 1, fill 2, …) and cap. Each zone gets its own ultrasonic beam, focused on the bevel geometry of that zone (bevel angle, depth).

The strength of this approach is that it maps the vertical position of a flaw directly onto a physical region of the weld cross-section. Imperfection types such as lack of penetration at the root or lack of fusion on the bevel face can then be classified by which zone channel responded. Contrary to a common misconception, zone discrimination gives more than position: the number of adjacent zone channels that respond, together with the way amplitude is distributed between neighboring zones, provides a first-order estimate of flaw height (through-wall height). In other words, even a pure zone discrimination system without TOFD can produce a height estimate as input to an engineering critical assessment (ECA). TOFD then confirms and sharpens that height measurement and covers the mid-wall volume of the weld through an independent channel. The reference practice for zone discrimination is ASTM E1961 (mechanized UT of girth welds using zonal discrimination with focused search units), and it is the common language of the pipeline sector.

Zone discrimination accuracy depends entirely on the WPS: if the weld preparation geometry changes, the zone definitions, the focal depths and the calibration block must all be reworked. Treat this as an essential variable of the AUT procedure.

3. Phased array probes and focal laws

Classic zone discrimination systems used a separate pair of fixed-angle focused probes for each zone. Bevel-face zones were typically covered by two probes in a tandem/pitch-catch arrangement, while the volume and the root/cap were scanned with pulse-echo probes. Modern setups instead program a single phased array (PA) probe with focal laws, generating the required angle and focal depth for each zone electronically. The system does this by applying software-controlled time delays across dozens of elements in the array.

The field advantages of the phased array approach are clear: one probe covers many zones, sectorial angle sweeps improve angular coverage of fusion-type flaws, and there is less probe-changing hardware to manage. A pipeline AUT band typically carries symmetrical PA probe pairs on either side of the weld, plus integrated TOFD (time-of-flight diffraction) channels to strengthen coverage of the weld center volume and flaw height determination. Running PA and TOFD together pairs the positional accuracy of zone discrimination with accurate height sizing. Automated PA technology on welds is framed by ISO 13588, and TOFD by ISO 10863.

4. Strip-chart and imaging output

The most familiar form of AUT output is the strip-chart: the horizontal axis is circumferential position around the weld (encoder clock position or angle), and the vertical axis is the channel for each zone. Signal amplitude in each channel is color- or grayscale-coded, and indications exceeding the defined evaluation threshold appear on a strip at their true position. On a single screen the inspector sees which zone produced an indication, and at which clock position, all the way around the circumference.

The strip-chart is accompanied by a coupling monitor channel. AUT data is only valid if acoustic coupling between the probe and the surface is maintained throughout the scan; wherever coupling is lost, the data is invalid and that section must be re-scanned. The TOFD B-scan image, the sectorial S-scan and, where needed, A-scan detail are presented alongside each other, which makes the evaluation multi-layered. All of these images are archived stamped with encoder position, and that gives AUT one of its most important advantages over RT: full traceability.

5. Calibration blocks and measurement

The reliability of AUT rests on calibration discipline. The calibration block used for zone discrimination is machined from a sample of the same material, the same pipe wall thickness, the same diameter and the same weld preparation geometry as the weld to be tested. Artificial reflectors representing the flaw type of each zone sit at the positions corresponding to that zone: usually angled flat-bottom holes (FBH) on the bevel face, surface notches for the root and cap, and side-drilled holes (SDH) for the volumetric reference.

During calibration, set the gain of each zone channel so that the relevant reference reflector produces a defined reference amplitude — for example 80% of full screen height (FSH) — and define the evaluation and recording thresholds relative to that reference. For TOFD, perform the time/depth calibration using the lateral wave and backwall signals. Verify the calibration (calibration check) at the start of each shift, on any probe or equipment change, and at the intervals defined in the procedure; if the deviation is unacceptable, re-evaluate every weld since the last valid check. The procedure states how temperature, coupling and scanning speed affect calibration validity. General UT techniques and testing levels follow ISO 17640 and ASME BPVC Section V, Article 4.

6. Speed, productivity and the advantage over RT

The value AUT brings to a pipeline spread falls under three headings:

Speed and productivity: A girth weld is scanned in a single pass, in a matter of minutes, and the evaluation follows almost immediately. RT drags a film exposure, processing and interpretation cycle behind it; AUT keeps pace with the spread and removes the bottleneck.
Working without radiation: With no ionizing radiation there is no exclusion zone, no work stoppage and no need for a night-shift window. Adjacent activities can run in parallel, which saves serious time on a project.
Permanent, digital records: Every weld is archived as position-stamped digital data. That data can be re-evaluated and audited later, and the flaw height measurements feed the ECA.

AUT is also more sensitive than RT to planar flaws — lack of fusion, lack of penetration, cracks. The difference comes down to orientation. RT makes the detection of a planar flaw dependent on its orientation relative to the beam. The flaw is only caught when its plane lies roughly parallel to the beam, producing a measurable difference in radiographic density along the beam path. When the flaw plane sits perpendicular or at an angle to the beam — as is common with tight, narrow lack of fusion and with cracks — no appreciable density difference forms and the flaw can be missed. Angle-beam UT is far less sensitive to orientation for these flaws, because the beam produces a strong echo from the flaw face. In exchange, AUT demands procedure and calibration discipline and qualified personnel; the setup threshold is higher than for RT. On short, one-off welds RT can still be the practical choice, but over long pipelines AUT wins by a wide margin.

7. The pipeline girth weld testing workflow

A typical AUT sequence in the field runs like this: define the zones and focal laws in the procedure according to the weld preparation and the WPS → set up and calibrate the system on the representative calibration block → mount the scanner on the guide band around the weld → acquire data in a single revolution → have the inspector evaluate the strip-chart, S-scan and TOFD data → mark the indication positions on the weld → make the accept/reject decision and report. Re-scanning is mandatory wherever coupling loss, calibration drift or a data gap is found.

8. Acceptance criteria: workmanship or ECA?

Pipeline AUT works with two different acceptance philosophies. Under workmanship criteria, indications are assessed against length and amplitude thresholds. Under the ECA / fitness-for-purpose approach, the height and length of the flaw are compared with fracture-mechanics-based acceptance envelopes, generated with BS 7910 methodology from material toughness, stress and service conditions. The ECA approach exploits the flaw-sizing power of AUT and reduces unnecessary repairs.

The principal sector documents for the welding, testing and acceptance of pipeline welds are API 1104, DNV-ST-F101 and ISO 13847. Fracture-mechanics-based alternative/ECA acceptance criteria are not in the body of API 1104 but specifically in Annex A (Alternative Acceptance Standards for Girth Welds); AUT itself is covered in the body. (Annex B of API 1104, which governs in-service welding, is unrelated to this subject.) The Annex A route requires, among other conditions, NDT of all welds and equal pipe wall thickness on the components being joined; without these conditions the alternative acceptance envelope cannot be used. Choosing the acceptance criteria and deriving the envelope is a joint responsibility of Level 3 and project engineering.

9. Common mistakes and verification

The problems seen most often in the field are: a calibration block that does not reflect the real weld geometry, an inconsistent circumferential position reference (0/12 o'clock) for the guide band, a coupling monitor channel that gets ignored, and zone/focal laws that are not updated after the WPS changes. Prove every AUT system before testing by procedure qualification on a verification block with blind flaws — preferably using production welds with known flaws and destructive verification (macro sections). That demonstrates the system really sees the flaw it is looking for. Where the acceptance approach is ECA, this qualification is indispensable.

From the field

On an offshore lay project, the mainline AUT strip-chart showed repeating low-amplitude indications in the root zone channel only, over a specific range of clock positions. The first reflex was "lack of fusion at the root." But the coupling monitor channel told a different story: the indications lined up exactly with the positions where the scanner had caught on weld spatter and briefly lost contact. We cleaned the surface, re-scanned the weld, and the indications were gone. The lesson is plain: AUT data is only real once coupling and encoder position are verified. When you read a strip-chart, question the validity of the data first and the flaw second. One more habit worth keeping: actually scan the calibration block at the start of every shift and confirm the reference amplitude — "it was set yesterday" has never supported an acceptance decision.

Worth stating once, plainly: every defect is a discontinuity; not every discontinuity is a defect — the acceptance criteria decide which is which.

Related standards

ASTM E1961-16(2021) — Mechanized UT of girth welds using zonal discrimination with focused search units (the zone discrimination reference).
API 1104, 22nd Edition (2021) — Welding of pipelines and related facilities; AUT in the body, ECA-based alternative acceptance in Annex A.
ISO 13588 — Ultrasonic testing of welds, use of automated phased array technology.
ISO 10863 — Ultrasonic testing of welds, TOFD technique.
ISO 17640 — Ultrasonic testing of welds; techniques, testing levels and assessment.
ASME BPVC 2023, Section V, Article 4 — Ultrasonic examination methods for welds.
DNV-ST-F101 — Submarine pipeline systems (testing and acceptance requirements).
ISO 13847 — Welding of pipelines.
BS 7910:2019 — Guide to methods for assessing the acceptability of flaws in metallic structures (ECA methodology).
ISO 9712:2021 — Qualification and certification of NDT personnel.

Take it to the field: To keep these NDT methods, standard references and field steps in your pocket — completely offline and free — take a look at the Doawise NDT Guide app.

At DoaWise we carry out automated ultrasonic testing of pipeline girth welds to ASTM E1961, API 1104, ISO 13588 and the related international standards. From zone discrimination to calibration verification, from TOFD sizing to ECA acceptance envelopes, we record every step as encoder-stamped digital data that can be archived and re-evaluated later. Our aim is not simply to say "accept" or "cut out" — it is to produce recordable, auditable results that engineering decisions can rely on with confidence.