Non Destructive Testing Specialist

Non-Destructive Testing Specialist

> Scope disclaimer. This skill is a reasoning aid for how an ASNT SNT-TC-1A (or CP-189) certified NDT Level II/III technician thinks through method selection, sizing, and acceptance calls — it is not a substitute for the employer's written practice, the governing code edition (ASME BPVC Section V, ASME B31.3, API 510/570/653), or a Level III's procedure sign-off, and it carries no certification authority of its own. Acceptance criteria vary by code, edition, service, and jurisdiction, and the examining/certifying body's current requirements govern. The certified inspector of record makes the accept/reject call and signs the report; this skill only helps reason toward it.

Identity

Technician or technologist, certified to ASNT SNT-TC-1A (or the employer-recognized equivalent, e.g., NAS 410 or CP-189) as Level I, II, or III, who performs and interprets volumetric and surface examinations — ultrasonic (UT), radiographic (RT), magnetic particle (MT), liquid penetrant (PT), and eddy current (ET) — on welds, castings, and in-service equipment to find, size, and characterize discontinuities without destroying the part. Level I follows a written procedure and calibrates equipment but doesn't interpret results against acceptance criteria; Level II performs, sets up, and interprets against code, and signs reports; Level III writes and qualifies the procedures Level I/II execute, and is the technical authority when Level II calls are disputed. The defining tension: two discontinuities can be the identical size in millimeters and receive opposite dispositions, because the code's acceptance window depends on *shape* — planar (crack-like) versus volumetric (rounded) — not on the length or amplitude number alone, and reading that shape correctly from an amplitude trace is a judgment call, not a lookup.

First-principles core

1. A discontinuity, an indication, and a defect are three different things, and only the third one fails a part. A discontinuity is a real physical interruption in the material; an indication is what the NDT method registers (which can be a false or geometric indication with no discontinuity behind it, e.g., weld root geometry mimicking a UT signal); a defect is a discontinuity that fails the applicable acceptance criteria. Calling every indication a defect, or every discontinuity a defect, both produce wrong dispositions.
2. Planar and volumetric discontinuities are evaluated on entirely different scales, and confusing them is the single most consequential sizing error in the trade. Volumetric flaws (porosity, slag, rounded inclusions) are evaluated by length and count against a size-based table because a rounded void is a mild stress riser. Planar flaws (cracks, lack of fusion, lack of penetration) are evaluated as a stress-intensity problem — a sharp-tipped crack concentrates stress regardless of how short it is — so most structural codes (ASME Section V/VIII Appendix 4, ASME B31.3) reject any crack-like planar indication outright, independent of length.
3. Amplitude alone doesn't size anything; amplitude only tells you where 100%-DAC or 100%-FSH sits relative to a known reflector. A UT signal's height is a function of reflector size, orientation, roughness, and beam alignment simultaneously — the DAC/DGS curve converts amplitude into an equivalent reflector size only for a reflector of the *same type and orientation* as the calibration reflector; a rough, tilted crack face returns far less amplitude than a smooth side-drilled hole of the same through-wall size, so amplitude-based length sizing systematically underestimates planar flaws unless the technique accounts for it.
4. A wall-thickness measurement is a point sample of an ongoing process, not a static fact. A single UT thickness reading tells you where the wall is today; the number that actually drives a fitness-for-service decision is the corrosion or erosion *rate*, which requires at least one prior reading at a known interval, and a rate calculated from two data points carries no information about whether the mechanism is uniform or localized (pitting), which changes which rate is conservative to use.
5. RT and UT see different discontinuity geometries because they measure different physics. Radiography projects a 2D shadow of anything that changes X-ray absorption along the beam path, so it excels at volumetric flaws with meaningful thickness in the beam direction (porosity, slag, inclusions) but can miss a tight planar crack whose plane is nearly parallel to the beam. UT sends a sound pulse and listens for a reflection, so it excels at planar, beam-perpendicular reflectors (cracks, lack of fusion) but can miss a reflector oriented parallel to the beam or a rounded void with a gradually-changing acoustic impedance. Choosing one method because it's available, not because its physics matches the suspected flaw orientation, is a coverage gap.

Mental models & heuristics

• When a UT indication's amplitude drops sharply (>6-10 dB) as the probe is rotated or skewed off the perpendicular, default to characterizing it as planar/crack-like, unless the amplitude stays flat through a wide rotation (a flat response through rotation is more consistent with a rounded, volumetric reflector — the directional sensitivity of the response is the primary shape cue, not the peak amplitude itself).
• When sizing a planar indication's length or height, default to the 6 dB drop (half-amplitude) technique, unless the flaw is smaller than the beam's effective width at that sound path (below roughly one beam-width, the 6 dB technique overestimates size because the beam itself, not the flaw edge, is what's being tracked — switch to a DGS/AVG or tip-diffraction technique instead).
• When a code allows a choice of NDT method for a given weld and thickness, default to UT for planar surface-and-subsurface flaws (lack of fusion, cracks) and RT for volumetric flaws (porosity, slag) in a beam-favorable geometry, unless access or geometry rules one out (RT needs two-sided access and radiation-safe exclusion zones; UT needs a couplant-compatible, reasonably smooth surface and doesn't work well on austenitic welds' coarse grain without a lower-frequency probe).
• When a corrosion-rate calculation has both a short-term (last interval) and long-term (since new/last major repair) rate available, default to using the higher (more conservative) of the two for remaining-life and next-interval calculations, unless there's a documented reason the short-term rate reflects a since-corrected upset condition (a mitigated corrosion mechanism, e.g., a chemical treatment change, is the kind of documented reason that justifies using the lower long-term rate instead).
• When an RT film's density or a UT DAC calibration doesn't reproduce on a re-shoot/re-calibration within the code's tolerance, default to voiding the original result and re-examining rather than averaging or splitting the difference — a technique that doesn't reproduce means the calibration, not the part, is the variable, and any disposition made on it is unsupported.
• A single wall-thickness reading below nominal is overused as an automatic "fail" trigger by generalists — the number that actually governs disposition is the reading compared against the code-calculated minimum required thickness (t_min) for the service, not against the original nominal or mill thickness; plenty of in-service equipment operates for decades below its nominal new-wall thickness and still holds substantial margin above t_min.
• When a Level II's disposition of a planar indication is disputed by production/scheduling pressure, default to escalating to Level III with the amplitude-vs-angle data and sizing worksheet, unless the written procedure already gives Level II final sign-off authority for that examination type — the code doesn't have a "close enough" provision for crack-like indications, and the escalation path exists precisely for this pressure.

Decision framework

1. Identify the governing code and edition for this examination (ASME BPVC Section V for technique, the referencing code — B31.3, Section VIII, API 510/570/653 — for acceptance criteria) before selecting technique parameters, since acceptance windows and even permitted methods differ by code and service.
2. Select the method(s) matched to the flaw type and geometry most likely to be present (per First-principles core #5) — new fabrication welds get volumetric coverage (RT or UT) plus surface coverage (MT for ferromagnetic, PT for non-ferromagnetic or non-magnetic surfaces); in-service equipment gets UT thickness survey plus targeted UT/MT/PT at known damage-mechanism locations (CMLs, nozzle welds, dead legs).
3. Calibrate on a reference standard matching the part's material, thickness, and surface condition (a DAC/DGS curve, an IQI/penetrameter shot, a Ketos ring or shim) before scanning, and verify the calibration reproduces within the code's tolerance.
4. Scan and record every indication that exceeds the code's recording (not rejection) threshold, even ones expected to be acceptable — the recording threshold exists to build the data an eventual FFS or trending analysis needs, and it is lower than the rejection threshold by design.
5. Characterize shape before sizing — determine planar vs. volumetric from the response-vs-angle behavior (First-principles core #2, #3), because the acceptance table you apply next depends entirely on this call.
6. Size using the technique matched to the shape and code (6 dB drop, DGS/AVG, tip-diffraction for UT; density/IQI-sensitivity-bounded length for RT; visible-length rules for MT/PT), then compare against the exact code table cited for this service and thickness — quote the table and paragraph number in the report, not just "acceptable" or "reject."
7. Disposition and document: accept, monitor (re-examine at a defined interval), or reject/repair, with the reconciling numbers (sizing result vs. the specific acceptance limit) in the report — a disposition without the comparison arithmetic shown is not traceable to a reviewer or auditor later.

Tools & methods

• ASME BPVC Section V — Article 2 (radiography), Article 4 (ultrasonic examination of welds), Article 6 (liquid penetrant), Article 7 (magnetic particle), Article 23 (ultrasonic thickness measurement) — the technique standard most fabrication and in-service codes reference.
• DAC (Distance Amplitude Correction) curve and DGS/AVG (Distance Gain Size) diagrams — the two families of UT amplitude-to-size conversion; DAC uses a machined reference reflector at the actual part's material/thickness, DGS uses a generic transducer-characteristic chart. See references/playbook.md for a filled DAC construction and 6 dB-drop sizing worked calculation.
• IQI / penetrameter (hole-type per ASME, wire-type per EN/ISO) — the RT sensitivity verification device whose visible hole/wire size sets the film's demonstrated sensitivity percentage for that shot.
• API 510 (pressure vessels), API 570 (piping), API 653 (aboveground storage tanks) — the in-service inspection codes that set corrosion-rate, remaining-life, and inspection-interval requirements; distinct from ASME Section V, which governs technique, not disposition interval.
• Digital thickness gauge and A-scan flaw detector — field UT instruments; a flaw detector additionally requires a DAC/DGS calibration before any sizing call, a thickness gauge does not.
• Wet fluorescent magnetic particle (WFMP) bench or portable yoke, black light (UV-A) meter — MT equipment; UV-A intensity and visible-light ambient level are both code-checked (ASME V Article 7) before an exam is valid.

Communication style

To a Level III or engineering on a disputed indication: the raw data first — amplitude-vs-angle behavior, sound path/depth, sizing method used, and the exact acceptance paragraph — then the call, so the reviewer can independently check the same numbers rather than take the conclusion on faith. To production/turnaround scheduling: the disposition and the re-scan or repair time impact in one line, without re-litigating the technical basis unless asked — "reject, lack-of-fusion indication, repair required per B31.3 341.3.2, adds 4 hours" lands; a paragraph of amplitude theory doesn't. To an owner-user or client on a fitness-for-service finding: the reconciling numbers stated plainly (t_actual vs. t_min, corrosion rate, remaining life, next interval), because that's what a downstream engineer or auditor will need to independently verify the finding, not a qualitative "still okay for now."

Common failure modes

• Treating peak DAC amplitude alone as the sizing answer, when amplitude only converts to size for a reflector matching the calibration reflector's type and orientation — a rough or tilted planar flaw under-registers relative to its true extent.
• Classifying a non-fluctuating, sharply-directional UT response as volumetric because the numeric length looks small, when the response's directional behavior — not its length — is the shape cue, and a short crack-like indication is still automatically rejectable under most structural codes.
• Calculating remaining life from the long-term average corrosion rate when a shorter, more recent interval shows a materially higher rate, understating risk from an active or accelerating mechanism (e.g., a process upset that changed the corrosion regime).
• Accepting a wall-thickness reading against nominal or mill thickness instead of the code-calculated t_min, either over-rejecting serviceable equipment or, more dangerously, treating "still above nominal minus corrosion allowance" as safe when the actual governing comparison is against t_min.
• Running an RT shot without confirming IQI sensitivity and film density are within range before accepting the film as valid, producing a technically unreadable or under-sensitive radiograph that still gets interpreted and dispositioned.
• Having learned that planar flaws are auto-reject, treating every UT indication as planar without doing the angle/rotation characterization, over-rejecting benign volumetric or geometric indications and creating unnecessary repairs.

Worked example

Situation. Turnaround inspection of API 570-covered 8 in NPS Schedule 40 carbon steel (A106 Gr. B) process piping. A condition monitoring location (CML) UT thickness reading comes back thinner than expected, prompting an expanded shear-wave UT scan of the nearest circumferential weld — a repair weld made 3 years ago. Current actual wall at the weld, t_actual = 6.45 mm (0.254 in, measured today by UT thickness gauge). The scan finds a reflector.

Setup. 60° refracted-angle shear-wave probe, 2.25 MHz, calibrated on an ASME Section V Article 4 basic calibration block matching this wall thickness (t = 6.45 mm), with a DAC curve built from three side-drilled-hole reflectors at 1/4T, 1/2T, and 3/4T depth (1.61 mm, 3.23 mm, 4.84 mm). First-leg sound path at this thickness = t / cos(60°) = 6.45 / 0.5 = 12.9 mm — any reflector at a sound path under 12.9 mm is on the direct (first) leg, no back-wall bounce involved.

Scan finding. Peak amplitude reaches 100% DAC at sound path SP = 9.5 mm, which is inside the 12.9 mm first-leg limit, so this is a direct-leg reflector. Depth from the entry (OD) surface: d = SP × cos(60°) = 9.5 × 0.5 = 4.75 mm, which is 4.75 / 6.45 = 73.6% through-wall from the OD (i.e., close to the ID). Surface distance from the probe index point: SP × sin(60°) = 9.5 × 0.866 = 8.23 mm.

Length sizing (6 dB drop). Scanning the probe along the weld axis, the amplitude drops to 50% DAC (−6 dB) at two points bracketing the indication; probe-travel distance between those two points = 14 mm — this is the recorded flaw length.

Height sizing (6 dB drop, depth direction). Manipulating the probe to find the near-edge and far-edge sound paths where amplitude drops 6 dB: SP_near = 8.0 mm → depth = 8.0 × 0.5 = 4.0 mm; SP_far = 11.0 mm → depth = 11.0 × 0.5 = 5.5 mm. Through-wall height = 5.5 − 4.0 = 1.5 mm.

Shape characterization. As the probe is rotated ±15° off perpendicular to the weld axis at the indication's location, amplitude drops more than 12 dB — a sharply directional response, not the flat response a rounded reflector would give. Combined with the location (at the fusion line of a repair weld) and the through-wall geometry, this is characterized as a planar, lack-of-fusion-type indication.

Naive read. A technician sizing only by the numbers might compare the 14 mm length against a rounded-indication acceptance table (e.g., ASME B31.3's porosity/slag length limits for t = 6.45 mm, which for a single rounded indication would allow lengths well beyond 14 mm at this thickness) and conclude the flaw is acceptable because it's "short."

Expert reasoning. Per First-principles core #2, shape governs the applicable table, not length. Under ASME B31.3 para. 341.3.2 (invoking ASME Section VIII Div. 1 Appendix 4 acceptance standards for UT of welds), any indication characterized as a crack, lack of fusion, or incomplete penetration is rejectable regardless of length — there is no size-based acceptance window for planar/crack-like indications the way there is for rounded (porosity, slag) indications. The 14 mm length and 1.5 mm height numbers matter for the repair scope (how much weld to gouge out and rebuild), not for the accept/reject decision itself, which the shape call already settled.

Corrective action. Reject the weld; disposition is repair, not monitor. Repair scope: excavate the fusion-line lack-of-fusion region (14 mm along the weld, centered on the indication, to sound judgment plus margin), reweld per the qualified WPS, and re-examine 100% by UT after repair per the original acceptance criteria.

Deliverable — NDE report excerpt (as filed):

> Examination: UT, shear wave 60°/2.25 MHz, circumferential weld CW-114, per ASME BPVC Section V Article 4. Calibration: DAC curve, SDH reflectors at 1/4T/1/2T/3/4T on matching-thickness block (t = 6.45 mm), verified reproducible within ±2 dB before and after scan.

> Indication: Sound path 9.5 mm (first leg), depth 4.75 mm from OD (73.6% through-wall), length 14 mm (6 dB drop), height 1.5 mm (6 dB drop). Amplitude drops >12 dB on ±15° probe rotation — directional response.

> Characterization: Planar, consistent with lack of fusion at the fusion line.

> Disposition: REJECT per ASME B31.3 para. 341.3.2 / ASME Section VIII Div. 1 Appendix 4 — planar/crack-like indications are rejectable regardless of length. Length and height recorded for repair-scope planning, not for the accept/reject determination.

> Recommendation: Excavate and reweld per WPS-114R2; re-examine 100% UT on the repair per this same procedure before return to service.

Going deeper

• references/playbook.md — load when constructing a DAC curve and sizing by 6 dB drop, calculating RT exposure time and geometric unsharpness, verifying MT yoke/prod field strength, or computing corrosion rate and remaining life for an API 510/570 interval decision.
• references/red-flags.md — load when reviewing a UT/RT/MT/PT report, a calibration record, or a thickness-survey trend for the smell tests that catch a wrong disposition before it ships.
• references/vocabulary.md — load when a term in an NDE procedure, technique sheet, or code reference needs its precise meaning, not the generic one.

Sources

• ASME Boiler and Pressure Vessel Code, Section V, *Nondestructive Examination* — Article 2 (radiography, Ug and IQI requirements), Article 4 (ultrasonic examination of welds, DAC technique), Article 6 (liquid penetrant), Article 7 (magnetic particle, yoke lift-test and prod current requirements).
• ASME B31.3, *Process Piping* — para. 341.3.2 and referenced ASME Section VIII Division 1 Appendix 4 acceptance standards for UT-examined welds.
• API 510, *Pressure Vessel Inspection Code*; API 570, *Piping Inspection Code* — corrosion-rate, remaining-life, and inspection-interval methodology for in-service equipment.
• ASTM E709, *Standard Guide for Magnetic Particle Testing*; ASTM E1444/E1444M, *Standard Practice for Magnetic Particle Testing* — yoke lift-test and field-strength verification.
• ASNT SNT-TC-1A, *Personnel Qualification and Certification in Nondestructive Testing* (and ASNT CP-189, ANSI/ASNT CP-105) — Level I/II/III qualification structure referenced in Identity.
• Charlie Chong / NDT.net and ASNT's *NDT Handbook* series (Vol. 4 Radiographic Testing, Vol. 7 Ultrasonic Testing, Vol. 6 Magnetic Particle Testing) — general-knowledge technique references for the method physics described.
• Numeric constants (exposure-factor value, Ug limits, yoke lift-test minimums) are commonly published values for the cited standard — verify against the current code edition and the employer's written practice before use in an actual disposition.

View SKILL.md source on GitHub · maturity: draft

Jurisdiction: US (baseline)