NDT Methods Compared: When to Use MPI, UT, PT or Radiography

An NDT scope written as "10% random NDT" is one of the most common specification failures in steel fabrication. It does not say which method, which joints, or what happens when something is found. The result is testing that satisfies a contract clause without meaningfully reducing risk.

Choosing well starts with a simple question: where is the defect likely to be — at the surface, or buried in the weld?

Surface methods

Magnetic Particle Inspection (MPI / MT)

A magnetic field is induced in the part and fine ferrous particles are applied. Where a discontinuity interrupts the field, the particles gather and make it visible.

• Finds: surface and slightly sub-surface defects — cracks, lack of fusion breaking the surface, laps, seams.
• Limits: ferromagnetic materials only. It will not work on austenitic stainless steel or aluminium. Sensitivity depends on defect orientation relative to the field, so fields are applied in two directions.
• Use it for: fillet welds, weld toes, fatigue-critical details, and anywhere a surface-breaking crack is the failure mode you fear.

Dye Penetrant Inspection (PT)

A penetrating liquid is applied, allowed to dwell, removed from the surface, and a developer draws trapped penetrant back out to reveal the flaw.

• Finds: surface-breaking defects only.
• Limits: truly surface-breaking only — it cannot see sub-surface flaws at all. Porous surfaces give false indications. Surface preparation and dwell time discipline drive reliability more than the operator's skill.
• Use it for: stainless steel and other non-magnetic materials, where MPI is unavailable.

Volumetric methods

Ultrasonic Testing (UT)

High-frequency sound is introduced into the material; reflections from discontinuities are timed and measured.

• Finds: internal defects — lack of fusion, incomplete penetration, cracks, laminations — and measures depth and through-wall extent.
• Limits: operator dependent in its conventional form. Flaws lying parallel to the sound beam can be missed. Requires access, couplant, and a suitable surface finish. Thin sections and complex geometry complicate interpretation.
• Use it for: full penetration butt welds in thicker sections; anywhere access is single-sided; as a safer alternative to radiography where radiation control is impractical. Phased array (PAUT) gives imaged, recordable results and largely answers the operator-dependency criticism.

Radiographic Testing (RT)

Radiation passes through the weld onto film or a digital detector, producing an image in which defects appear as density changes.

• Finds: internal volumetric defects — porosity, slag inclusions, incomplete penetration — with an objective, permanent, reviewable record.
• Limits: poor at detecting tight planar defects such as lack of fusion lying perpendicular to the beam. Requires two-sided access. Radiation safety means cordoning areas and, usually, night shifts — a real schedule cost. Sensitivity falls off in very thick sections.
• Use it for: pipe butt welds and pressure-retaining joints, and where the client requires a permanent image.

Choosing between UT and RT

The honest summary: RT gives you a picture, UT gives you depth. For a planar defect like lack of sidewall fusion — the defect most likely to fail a structural joint — UT is generally the more reliable detector, while RT excels at volumetric defects and produces evidence a third party can re-examine years later.

Many modern specifications now accept or prefer PAUT in place of RT for thick butt welds, because it produces a recordable image without the radiation control burden.

Writing an NDT scope that works

1. Tie the extent to the execution class and joint criticality, not to a flat percentage. Full penetration butt welds in primary members are not the same risk as fillet welds on secondary bracing.
2. Name the method per joint type in the inspection and test plan.
3. Name the acceptance standard and level — for example ISO 5817 quality level B, C or D, or the relevant AWS/AISC criteria. "Free from defects" is not an acceptance criterion.
4. Specify technician certification — typically ISO 9712 or SNT-TC-1A Level II for interpretation, with Level III for procedure approval.
5. Define the escalation rule. If a random sample fails, does the sample double? Does 100% of that welder's work in the period get tested? Agree this before anyone finds anything.
6. Require visual inspection first. Competent visual inspection by a qualified welding inspector catches more real defects per dirham than any other method, and it is the one most often skipped.

A note on timing

Certain defects — hydrogen-induced cold cracking in particular — can appear hours after welding. For higher strength steels and thicker sections, specifications commonly require a delay, often 48 hours, before final NDT. Testing too early produces a clean report on a joint that cracks later.

Al Saqr Al Ahmar works to approved inspection and test plans with third-party NDT on structural steel and pipeline scopes across the UAE. Read how welding procedures are qualified, or discuss your inspection requirements.