Who Approves Your UT Procedures in the Permian Basin

What a UT Level III actually approves

Four things carry a UT Level III's signature. The procedure: the controlling document stating equipment, probes, frequencies, wedge angles, calibration basis, scanning coverage, recording level, evaluation rules and acceptance criteria against a named code. The technique sheets written from it for a specific geometry, thickness and material. The qualification material — the general and specific examinations, and the practical specimens with their characterised flaw content. And the certifications themselves.

The procedure is where most of the technical decision-making lives, and it is where inherited documents cause the most trouble. A procedure copied from another company reflects that company's equipment, its calibration blocks and its acceptance basis. Run it with different probes on different material and the numbers it produces are not the numbers it was validated to produce. The Level III's job is to make it match the shop that will actually use it.

The examination material is the part shops underestimate. A practical examination in UT is not a conversation about theory; it is a candidate finding, locating and sizing reflectors in a specimen whose flaw content has been independently characterised, against a marking scheme that produces a grade. Without characterised specimens and a key, there is no practical examination — there is a supervisor's opinion, recorded on a form.

Straight beam, angle beam, and what certification actually covers

A UT certification names a method, but the work inside that method is not uniform. Straight beam, normal incidence work — thickness measurement, lamination checks, plate and forging examination — depends on velocity calibration, zero offset, coupling and interpretation of back-wall behaviour. Angle beam work depends on refracted angle verification, skip distance geometry, surface distance and depth arithmetic, and beam path attenuation. A technician can be entirely reliable in one and unreliable in the other.

The practice should say where the boundaries are. Some shops certify UT as a single scope and rely on the practical to cover both; others define applications explicitly — thickness measurement, weld examination, corrosion mapping, tubular inspection — and certify against each. The second approach is more work and dramatically easier to defend, because the auditor's question is always the same: what did this individual demonstrate, and on what specimen?

The boundary that catches most companies is encoded and mechanised work. Phased array and TOFD require competence in setup and data interpretation that manual A-scan practice does not build. A UT Level II who has never constructed a focal law or verified an encoder is not qualified to produce phased array data merely because the equipment is nominally an ultrasonic instrument.

DAC, DGS and the transfer correction nobody documents

A distance-amplitude correction curve is built by responding to identical reference reflectors — typically side-drilled holes in a basic calibration block under ASME Section V Article 4 — at increasing sound paths, and joining the peaks. Everything after that is referenced to the curve. The assumption buried inside it is that the block and the component attenuate sound the same way. They frequently do not.

Surface condition, grain structure, coating, curvature and temperature all change the energy that gets into and back out of the part. If the component is more attenuating than the block and no transfer correction is applied, every indication reads low relative to the curve, and marginal reflectors sit below the recording level. The correction is straightforward to determine and simple to state in a procedure. It is routinely absent, and its absence produces clean reports on parts that are not clean.

DGS avoids the block set by using published probe diagrams to relate amplitude, distance and equivalent reflector size, which is efficient and dependent on the probe matching the diagram. Substitute a probe of different element size or damping and the relationship no longer holds. Structural work under AWS D1.1 adds a third scheme entirely, where the indication rating subtracts a reference level and an attenuation term derived from sound path — so an error in measuring sound path moves the rating and can move the weld across an acceptance boundary. The Level III chooses one basis per procedure, states the compensation rules, and qualifies technicians against the basis they will actually use.

The thickness arithmetic that produces non-conservative readings

Ultrasonic thickness measurement looks like the simplest task in the method and generates the most damaging errors, because the instrument always returns a number and the number always looks authoritative. Three arithmetic traps recur across Permian work.

Velocity mismatch: a gauge calibrated on carbon steel at roughly 0.233 inches per microsecond and then used on austenitic stainless at roughly 0.223 reports about four and a half percent more metal than exists — on a half-inch nominal wall, about twenty-two thousandths of an inch that is not there. Temperature: velocity in steel falls approximately one percent per hundred degrees Fahrenheit, so an uncompensated reading on hot service overstates remaining wall, again in the unsafe direction. Doubling: on thin sections, a gauge can lock onto a second back-wall echo and report twice the true thickness, which is the worst possible failure mode because it appears exactly where the wall is thinnest.

Then the interpretation traps. A flat probe on a pitted internal surface reads the shortest path it can find or none at all, so pit depth is systematically misrepresented. Averaging rules matter too: minimum thickness assessment in API 570 and API 653 depends on how a locally thinned area is treated and over what length readings may be averaged, and a technician applying an averaging rule the procedure never stated is inventing engineering judgement. The Level III writes those rules down, and writes down when a reading must be escalated rather than averaged away.

Flaw sizing, and what it is actually used for locally

Detection and sizing are different competencies. Detection asks whether something is there; sizing asks how deep and how long, and it feeds decisions with real money attached — accept or repair, run or shut down, and in some cases an engineering critical assessment or a fitness-for-service evaluation under API 579-1/ASME FFS-1.

In the Permian, sizing shows up in a few identifiable places. Pipeline girth welds where API 1104's alternative acceptance criteria are used instead of workmanship criteria: those criteria depend on flaw height, and flaw height depends on a sizing technique the Level III has qualified, not on an operator's estimate. Shop-built pressure vessels where a rejected weld means grinding out and re-welding a heavy section, and the difference between an eight-millimetre and a fourteen-millimetre indication is days of schedule. In-service equipment where a crack-like indication triggers an assessment that will consume engineering time.

The distinction the Level III has to enforce is between amplitude-based evaluation and true height measurement. Amplitude tells you a reflector responds like a reference reflector; it does not tell you the reflector's through-wall extent. A procedure that assigns height from amplitude, and an acceptance criterion that requires actual height, are not compatible, and the mismatch is usually invisible until the result is contested.

Encoded phased array and TOFD: the qualification nobody budgets for

Encoded techniques change the economics of ultrasonic inspection and change the qualification burden with them. Phased array produces an imaged, recorded dataset that a third party can re-examine months later, which is an enormous advantage — and it means errors are permanent and reviewable rather than transient. The setup carries the risk: focal law construction, wedge and element selection, velocity and wedge delay calibration, sensitivity calibration across the sweep, encoder resolution, and scan plan coverage.

Verification is the step that gets skipped. A scan plan is a prediction about where sound goes; it is not evidence. The Level III's requirement is a demonstration on a representative specimen with known reflectors at the depths and orientations the technique claims to cover, documented well enough that an auditor can see the technique found what it was supposed to find. Where ultrasonic examination substitutes for radiography under Section VIII Division 1, this demonstration is not optional and the Authorized Inspector will ask for it.

TOFD brings its own boundary condition. Its accuracy in through-wall sizing is excellent through the body of the weld and poor near the scanning surface, where the lateral wave creates a dead zone. A procedure that relies on TOFD alone and does not specify a complementary technique for the near surface has an uncovered region — and near-surface flaws are among the most significant in cyclically loaded service.

Who buys ultrasonic authority around Midland, and why

The demand is not evenly distributed. Fabrication shops building separators, heater treaters, free-water knockouts and skid packages to ASME Section VIII Division 1 need weld examination procedures, qualified technicians and, increasingly, the ability to offer ultrasonic examination in place of radiography to avoid shutting a bay down for a shot. Pipe spool shops feeding gathering, transfer and plant piping need the same, plus the throughput to make it viable.

In-service work is the other half. API 570 condition monitoring on gathering and transfer systems across the Midland and Delaware sub-basins. API 653 tank inspection on the battery tanks and larger terminal storage. Cryogenic gas processing facilities run by the major Permian midstream operators, where exchanger, cold-box and high-pressure piping examinations carry both mechanical integrity and process safety weight. Corrosion under insulation surveys on hot service, where the ultrasonic answer is only as good as the temperature compensation behind it.

Then the specialised corners. OCTG threading and inspection yards in the Midland–Odessa corridor, running automated wall and flaw channels at production rates. Refinery turnaround work drawing contractors toward Big Spring. And a small amount of aviation-related maintenance activity at Midland International Air and Space Port, which sits under a completely different personnel standard in NAS 410 and should never be absorbed into an oilfield written practice by convenience.

Scope traps in UT certification

Limited certification is a legitimate tool and a frequent liability. Certifying a technician for thickness measurement only, or for one product form, is defensible when the practice states the boundary and the reports respect it. The failure mode is predictable: the limited technician is on site, an anomaly appears that falls outside their scope, and they evaluate it anyway because the alternative is a delay. The practice needs an escalation rule, not just a boundary.

Interruption of activity is the second trap. Ultrasonic interpretation degrades with disuse faster than most methods, and Permian schedules routinely park a technician on other work for a year. If the practice is silent on interruption, nobody triggers a reassessment and the technician returns to a method they have not exercised — with a certificate that is still nominally valid.

The third is inherited procedures. A procedure that arrived with a previous Level III, or with an acquired business, describes equipment and blocks that may no longer exist. Reviewing them is unglamorous and it is the highest-yield hour a UT Level III can spend in a new engagement. Atlantis provides UT procedure development and qualification, technique demonstration, examination and practical material, certification within your written practice, and independent review of contested ultrasonic data — scoped to your codes and equipment, with a quote provided on request via info@atlantisndt.com.

Request a consultation