Encoded PAUT/AUT holds a near-flat 5–10 minutes per weld across pipeline wall thicknesses, while radiography needs longer exposures and heavier sources as steel in the beam grows past the practical Ir-192 range of about 2.5" of penetrated steel [S3]. On heavy-wall mainline and tie-in welds, that single fact usually decides the inspection method before the per-joint dollar number is even drawn up.
For thinner small-bore work the trade shifts: radiography's footprint stays manageable, and the cost gap to PAUT narrows. The per-joint comparison is therefore not a single ratio but a decision driven by wall thickness, weld count, code acceptance path, and how many trades the exclusion zone will stall [S3][S5].
What actually moves the per-weld cost
Three cost drivers dominate the PAUT-versus-RT decision on pipelines. The first is the instrument and scanner capital: the phased array unit, encoder, and scanner hardware are priced well above a conventional UT set, and on top of that the technician must hold a CAN/CGSB Level 2 UT plus a PA-UT endorsement rather than just Level 2 UT [S2]. The second driver is the procedure qualification, which is more complex for PAUT and needs a demonstration block before production scanning can start. The third driver is the operational footprint: an Ir-192 shoot can require a 30 to 100 foot exclusion zone that pushes pipefitters, scaffold builders, and insulators off the unit until the shoot, the clear-out, and the film processing are done, and that lost-time cost is where PAUT usually wins on a running plant or stacked turnaround [S5].
PAUT also avoids film processing chemistry and produces an encoded digital record that a Level III reviewer can re-read later, which removes the "what did the inspector actually see" gap that a written UT or RT record leaves behind [S2][S4]. The cost of generating that record is real, but it is paid once at setup and is amortised over the whole weld count on a campaign.
Where radiography still earns the job
Thin-wall small-bore piping and complex castings remain RT's strong suit. Below the Ir-192 ceiling near 2.5" of penetrated steel, exposure times are short, sensitivity is good, and the geometric unsharpness is manageable with a single-wall crawler shot, which is why a 48" pipe run is typically shot as double-wall (beam crosses both walls, twice the wall thickness, poorer sensitivity than single-wall crawler) rather than crawler [S3]. Past that ceiling the work moves to Co-60 or a linear accelerator, and the cost-per-shoot curve bends up sharply.
There are also inspection scenarios where RT is the only volumetric method the owner or the code will accept without a long equivalency argument. ASME Section V and Section VIII, API 1104, CSA Z662 Annex N, and ISO 13588 all name PAUT explicitly, but acceptance is on a project-by-project basis, and a procurement specification written around RT for legacy reasons still forces the truck out [S2][S4]. On populated units, inside pipe racks, near control rooms, on offshore platforms, and in occupied process buildings, however, the safety overhead of the exclusion zone usually drives the switch to PAUT regardless of what the original spec said [S5].
Code acceptance and procedure qualification

PAUT is accepted in lieu of RT for new pipeline girth welds under API 1104, CSA Z662 Annex N, and ASME Code Cases 2553 and 2557, with ISO 13588 covering the application of PAUT for weld inspection in general [S2][S4]. Acceptance under AWS D1.1 for structural welding is conditional on meeting the acceptance criteria in the code. The pipeline-codes path is the one that matters most for mainline and tie-in work, and it is mature enough that most owners will accept a PAUT procedure qualification in place of an RT one without an equivalency submission.
For pre-service inspections, cost reduction is typically given more weight than marginal performance improvement, on the rationale that the inspection is acceptable as long as the applicable code is satisfied [S4]. That framing is why the per-joint dollar number, not the flaw-detection physics, usually drives the method choice on a build, and why a credible per-joint comparison has to include the procedure qualification effort, not just the scan time on the day.
Decision matrix: pick the method by job shape
For pipeline girth welds on wall thickness above roughly 0.5" with weld counts in the dozens to hundreds, PAUT wins on scan time (5–10 min per weld, near-flat with wall thickness), on operational footprint (no exclusion zone, adjacent crews keep working), and on record quality (encoded, replayable, Level III reviewable) [S3][S5]. For thin-wall small-bore shop work where the weld count is low and the geometry is complex, RT is often cheaper once exposure and film-processing are priced in. For heavy-wall refinery and petrochemical nozzles inside a turnaround with stacked trades, PAUT's exclusion-zone savings dominate the calculation. For work in a populated process building or on an offshore platform, PAUT is essentially the only viable volumetric method because the RT exclusion zone cannot be set up safely [S4][S5].
For the all-in per-joint cost on a pipeline campaign, the honest answer is that the number sits in a range rather than at a single point, and it moves with weld count, wall thickness, crew stacking, and whether the inspection has to be done in-service. Headline percentages claiming PAUT is some fixed amount cheaper than RT deserve skepticism, but the underlying drivers (scan time scaling, exclusion-zone labour, encoded record) are real and quantifiable on a per-job basis [S5].
Spec-side guardrails before you sign the ITP

Three items need to be locked down before a PAUT-for-RT substitution is written into the Inspection and Test Plan. First, the code path: name API 1104, CSA Z662 Annex N, or the relevant ASME Code Case, not a generic "PAUT in lieu of RT" clause [S2]. Second, the qualification: require a demonstration block scan on representative joint geometry before production starts, and require the technician card to be Level 2 UT plus PA-UT endorsement, not just Level 2 UT [S2]. Third, the data deliverable: require encoded scan files (S-scan plus B/C where used) to be archived with the job, indexed by weld map, and re-readable by a Level III who was not on the job, so the data survives crew turnover and incident review [S2][S4].
On the RT side, if the method is kept for legacy or thin-wall reasons, the equivalent guardrails are IQI sensitivity verification, source-to-film distance discipline, and the same per-weld indexing of the digital or film record, which is where the adjacent decision on film versus digital radiography tends to surface as the second-tier cost question on the same job.
Where the per-joint number will move next
Two signals are worth tracking through 2026. First, more oil and gas owners are adding PAUT to their approved vendor lists for new-build pipeline girth welds, which compresses procedure qualification effort and pulls the PAUT per-joint number down on repeat campaigns [S5]. Second, the practical Ir-192 ceiling near 2.5" of penetrated steel is the line at which RT cost curves bend upward as the work moves to Co-60 or linear accelerator, and any tightening of site radiation-safety rules widens that gap in PAUT's favour on heavy-wall work [S3]. For procurement engineers, the actionable next step is to ask vendors for a per-joint number broken out by setup, scan, qualification, and crew-time lost to exclusion, then to weigh that against the same breakdown for RT on the specific wall-thickness and weld-count profile of the project.