NDT Method Selection: Which Method for Which Flaw?

In non-destructive testing (NDT), the most expensive mistake is not a misread result — it is choosing the wrong method. A test performed flawlessly with a method that cannot see the flaw looks "clean" on the report, yet sends a flawed part into service. That is why an experienced Level III does not start by picking up a probe. The work starts with the right questions. Is the discontinuity I am looking for at the surface or below it? Is the material ferromagnetic, conductive, radiation-transparent? Which way is the flaw likely to run? From how many surfaces can I reach the part? Do I need to record, archive and audit the result? Below I work through that decision chain criterion by criterion, then compare the eight core methods with their strengths and weaknesses, and close with a selection matrix you can actually use.

1. Two axes: surface or internal, planar or volumetric?

NDT method selection does not rest on one question but on two complementary axes. The first axis is location: is the discontinuity surface-breaking (or near-surface), or is it inside the material — embedded? The second axis is morphology: is the discontinuity planar, or volumetric (rounded)? Together, these two axes rule out most methods before you ever leave the office. A common mistake is to assume "internal flaw = volumetric flaw." An internal crack is embedded, but it is planar — and that distinction drives the method choice directly.

For surface and near-surface discontinuities (cracks, overlap, grinding cracks, surface-breaking porosity), the primary tools are VT, PT and MT, plus ET on conductive materials. These methods do not "look into" the volume. Their strength is catching the opening at the surface, or the magnetic/electrical disturbance just beneath it.

For internal or embedded discontinuities, RT, UT, PAUT and TOFD take over. These methods scan the volume with energy that travels through the material — X-rays, gamma rays or sound. Here the second axis becomes critical, and it splits internal flaws in two:

Volumetric embedded flaws (internal porosity, slag, casting cavities) attenuate the beam similarly from every direction. RT handles these comfortably and gives a clear image.
Planar embedded flaws (lack of fusion, internal or edge cracks, lamination) are tight, orientation-dependent discontinuities. Here UT/PAUT leads, while RT — being extremely orientation-sensitive — can easily miss them.

In practice, most weld testing is two-layered: a surface method (MT or PT) first for surface flaws, then a volumetric method (RT or UT/PAUT) to screen out internal ones. The expectation that "one method sees everything" is the most widespread conceptual error in NDT.

2. Material: ferromagnetic, conductive, radiation-transparent?

Some methods depend on a physical property of the material. No property, no method.

MT works only on ferromagnetic materials (carbon steel, low-alloy steel, ferritic stainless). Austenitic stainless (304/316), aluminum, copper and titanium cannot be magnetized. On those materials, use PT for surface flaws, not MT.
ET requires a conductive material. It applies to both ferromagnetic and non-ferromagnetic conductors; ferromagnetic materials may need magnetic saturation. It is the primary method for heat exchanger tubes (non-ferromagnetic), aluminum structures and conductive coating checks.
PT is material-independent, but it needs a non-porous, non-absorbent surface. Spongy castings or porous ceramics produce false indications.
For RT the governing parameters are density and thickness. Dense or thick sections demand higher energy and output (kV and mA for X-rays; source activity for gamma — Ir-192 or Co-60) and longer exposure times. On very dense materials such as lead, transmission drops sharply.
UT depends on sound transmission. It performs excellently in fine-grained steel, but in coarse-grained castings and austenitic welds, sound scattering (structural noise) can drown the signal. There you need lower frequency and often PAUT/TOFD or special angle probes.

3. Flaw orientation and morphology

Here is the fact most often forgotten in NDT: a method's probability of detection depends on how the flaw is oriented. Match the method to the geometry of the flaw you expect.

Planar flaws (cracks, lack of fusion, sidewall lack of fusion) are the most dangerous discontinuities, because their notch effect initiates cracking. UT/PAUT catches them very well — as long as the sound beam strikes the flaw plane perpendicularly. With RT the logic is reversed: RT can see a planar crack only when the beam passes parallel to the crack plane (or very close to it, typically within ±5–10°). When the beam hits the crack plane perpendicularly, the tight gap creates too little density difference and the crack is missed. So RT is extremely orientation-sensitive to planar flaws, and behaves exactly opposite to UT.
Volumetric, rounded flaws (porosity, slag, casting cavities) present the same missing material to the beam from any direction, which makes them RT's strongest flaw class: sharp radiographs, easy interpretation.
In MT, flaw orientation relative to the magnetic field is critical: a crack gives the strongest indication when it lies perpendicular to the magnetic flux lines, and stays invisible when parallel. That is why MT is always performed with magnetization in two perpendicular directions.
ET is strong on cracks that open perpendicular to the surface; it detects the flaw that interrupts the eddy current path.

In short: if you are hunting porosity or slag, RT is comfortable; if you are hunting cracks or lack of fusion, UT/PAUT and directional surface methods take the lead.

4. Access, geometry and surface condition

Even when the physics works, how you reach the part decides the method.

RT requires double-sided access: the radiation source (X-ray tube or gamma isotope) on one surface, film or detector on the opposite one. On piping, solid shafts or closed volumes accessible from one side only, that may be impossible; panoramic and double-wall techniques have their limits.
UT/PAUT/TOFD can work from one side — a major advantage on pipelines, pressure vessels and solid sections.
Surface roughness directly affects couplant performance in UT and MT/PT sensitivity; a rough casting surface will not give reliable results until it is ground.
Temperature: standard probes and penetrants are valid only within a defined temperature range. High-temperature work needs special couplant and probes, or a high-temperature penetrant.
When the geometry is complex (threads, radii, thin walls), beam spread and echo interpretation get harder; sometimes VT plus PT is the only practical option left.

Radiation safety is an access constraint too. The area must be cleared during RT, which drives plant shift and shutdown planning and often makes UT the operationally attractive choice.

5. Speed, productivity and cost

When several methods are technically suitable, the decision shifts to productivity and total cost.

VT is the fastest and cheapest method. It opens every NDT job and screens out most flaws before any other method is set up.
PT and MT have low equipment cost, but PT runs a slow cycle: pre-cleaning, penetrant application, dwell, developer, evaluation.
RT's hidden cost is film and processing, radiation safety, area clearance and exposure time. Digital radiography (CR/DR) cuts that cost but raises the equipment investment.
UT gives fast, immediate results, with high throughput in the hands of a competent operator.
PAUT carries the highest up-front investment and training requirement, but it is the most productive volumetric method in terms of scanning speed and coverage. That is the main reason it has been displacing RT on large weld volumes.

The right question is not "which one is cheap?" but "what is the total cost per unit of reliable testing?" A method that looks cheap can turn out to be the most expensive one, once you price the flaw it missed in the field.

6. Recording, traceability and reportability

Some methods produce a permanent, auditable record by their very nature; others do not. Combined with code or client requirements, this alone can decide the selection.

RT has historically been the "gold record": film or digital image is a permanent document of the moment of testing, and can be re-evaluated years later.
PAUT and TOFD produce position-linked, archivable data when scanning is encoded. Sectorial and amplitude images and time-of-flight records remain open to review. This is the second reason PAUT and TOFD are a strong alternative to RT in modern weld testing.
Conventional manual UT, by contrast, relies largely on the operator's real-time interpretation. Its record is the report the operator fills in — without encoded scanning, re-evaluation is limited.
VT, PT and MT are usually documented with a report plus photographs; the indication itself is not stored as a permanent signal.

If the contract calls for "recorded, traceable testing," then even when two methods satisfy the same physics, the one that produces a record (RT, or encoded PAUT/TOFD) wins.

7. Strengths and limitations of the methods

VT (visual testing): The most basic and most widely used method. Cheap, fast, applicable at almost every stage. It sees only the surface and surface-breaking flaws; it depends on lighting, access and the inspector's eye. It is the mandatory first step of every NDT program.
PT (liquid penetrant testing): Material-independent, with high sensitivity on surface-breaking cracks and porosity. Simple equipment. It sees only surface-breaking flaws; it gives false indications on porous surfaces and misses flaws whenever cleaning is not meticulous; slow cycle.
MT (magnetic particle testing): Fast and highly sensitive on surface and near-surface flaws in ferromagnetic material. Limited to ferromagnetic materials; detection depends on flaw direction relative to the magnetization (two directions are mandatory); demagnetization may be required.
ET (eddy current testing): Fast on surface and near-surface cracks in conductive material, tube testing, and coating or conductivity measurement; needs no couplant (dry) and produces a record. Interpretation requires experience; conductive materials only; penetration depth is limited; sensitive to geometry and lift-off effects.
RT (radiographic testing): Volumetric; superior on rounded flaws such as porosity and slag, produces a permanent image, tolerant of geometry. Its weaknesses are radiation safety, the need for double-sided access, the risk of missing a planar crack when the beam meets the crack plane perpendicularly rather than parallel to it, and long exposure times on thick sections.
UT (ultrasonic testing): Volumetric; strong on planar flaws and cracks and on thickness measurement, works from one side, immediate, no radiation. Heavily dependent on calibration and operator competence; struggles in coarse-grained and austenitic material; weak record in its conventional form.
PAUT (phased array ultrasonic testing): All the strengths of UT plus angular scanning, fast coverage, encoded records and sectorial imaging. High investment and advanced training; calibration and setup are complex.
TOFD: Very accurate in flaw sizing and height measurement, fast, and it produces a record. It covers most of the through-thickness volume; its limits are the near-surface dead zone and the back-wall dead zone, interpretation that demands expertise, and reliance on transit time rather than amplitude. For that reason it is used together with pulse-echo or PAUT.

8. Practical selection matrix

The pairings below are a practical starting point for typical carbon steel weld and pressure equipment scenarios. The final selection always follows the construction or in-service code:

Surface crack in a carbon steel weld → MT (primary); PT where MT is not available.
Surface crack in an austenitic stainless weld → PT (MT will not work).
Internal porosity or slag in a weld → RT (clearest interpretation).
Lack of fusion or internal crack in a weld → UT/PAUT (RT may miss it).
Heavy wall, large weld volume, record required → PAUT + TOFD (accepted in place of RT under ASME Section VIII, Division 1 subject to specific conditions — the UT-in-lieu-of-RT provisions, which impose thickness limits and additional documentation).
Pipe wall thickness / vessel shell thickness and corrosion monitoring → UT thickness measurement and corrosion mapping.
Heat exchanger tubes → ET (non-ferromagnetic), RFT or IRIS (ferritic and special cases).
Surface crack in an aluminum aircraft structure → PT or ET.
Internal cavity in a casting → RT (strong on rounded flaws).
First step in every case → VT.

The matrix is a starting point, not a recipe. The real selection emerges at the intersection of flaw type, material, access, code and record requirement — and it usually ends in the layered use of more than one method.

9. Related standards

Read the references below against the current edition in force for the method in question. These standards are revised regularly (EN ISO 17636-1, for example, moved from 2013 to 2022), and the test plan must always state the year and edition it relies on.

ISO 9712: Qualification and certification of NDT personnel (Level I/II/III); on the US side, ASNT SNT-TC-1A / CP-189.
ASME BPVC Section V: NDT methods — Article 2 (RT), Article 4 (UT/PAUT/TOFD), Article 6 (PT), Article 7 (MT), Article 8 (ET), Article 9 (VT).
EN ISO 17635: Non-destructive testing of welds — general rules for method selection (metallic materials).
VT: EN ISO 17637 (visual testing of welds).
PT: EN ISO 3452-1, ASTM E165/E1417; acceptance: EN ISO 23277.
MT: EN ISO 9934, EN ISO 17638, ASTM E709/E1444; acceptance: EN ISO 23278.
ET: EN ISO 15549 (general principles), ASTM E243 (copper and copper alloy tubes).
RT: EN ISO 17636-1 (film) / -2 (digital); acceptance: EN ISO 10675-1 (steel).
UT: EN ISO 16810 (general), EN ISO 17640 (welds); acceptance: EN ISO 11666.
PAUT: EN ISO 13588; acceptance: EN ISO 19285.
TOFD: EN ISO 10863; acceptance: EN ISO 15626.

On pressure equipment built and examined to ASME, acceptance follows the ASME Section VIII appendices directly, not EN ISO: RT acceptance → UW-51 (100% RT) / UW-52 (spot RT); UT acceptance → Mandatory Appendix 12 (with Article 4 and the related conditional provisions for UT in lieu of RT); MT acceptance → Mandatory Appendix 6; PT acceptance → Mandatory Appendix 8. Acceptance criteria come from the construction or in-service code the examination is bound to (for example ASME Section VIII, the applicable piping or welding code, or API standards) — and method selection is finalized within the frame of that code and the examination plan.

From the field

What derails method selection on site is usually not physics — it is conditions. A shutdown where you cannot clear the area for RT, a coarse-grained austenitic weld that fights UT, or one line in a client spec — "all testing shall be recorded" — can rule out the technically ideal method in an instant. Experienced inspectors never fall in love with a single method. They put flaw type, material, access, code and record requirement on the same table, and usually end up with a layered plan: VT first, then MT or PT for the surface, then UT/PAUT or RT for the volume. The most common mistake is pointing whatever instrument happens to be in the van at every flaw. The right method is the one with the highest probability of detection for that flaw, and that call belongs to the test plan, not to the equipment on hand.

This article is for educational purposes and does not replace the official standard, the code or the written test procedure — the final decision always rests with the code in force and the responsible inspector.

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 non-destructive testing services to international standards. For every job we select the right method according to flaw type, material and access conditions, and we build layered test plans covering VT/PT/MT/ET/RT/UT/PAUT/TOFD. Our aim is not only to find the flaw, but to support plant operators in making safe, planned maintenance decisions with recordable, auditable and re-evaluable results.