Buying Weld Inspection for US Fabrication and Field Construction

What Weld Inspection Establishes About a Joint

A weld inspection scope exists to establish four things about a joint, and every method in the stack maps to one of them. That the weld was made to a qualified procedure by a qualified welder. That its external geometry, meaning size, profile, undercut and reinforcement, meets the code. That its surface is free of cracks and porosity that break through. And that its internal volume contains no discontinuity larger than the code allows. A scope that covers three of the four is not a cheaper inspection, it is an incomplete one.

Most rejectable conditions on US fabrication are caught in the first two. Wrong electrode, missing preheat, undersized fillet, arc strikes, unfilled craters, profile that violates the code: none of these need a volumetric method, and none of them survive a competent visual inspector working during production rather than after it. Buying volumetric NDT while leaving visual inspection to the fabricator's own foreman inverts where the value is, and it is the most common structural mistake in weld inspection scoping.

The volumetric methods earn their keep on the joints where failure is expensive and the defect is invisible: full-penetration groove welds in structural connections, pressure-retaining seams, and pipeline girth welds. Those are the joints the codes single out for volumetric examination, and they are where an independent inspection contract pays for itself several times over on a single catch.

Choosing the Method by Joint and Defect Type

Method selection is not a preference. Magnetic particle detects surface and slightly subsurface discontinuities in ferromagnetic material, and it is the fastest, cheapest and most sensitive way to find a toe crack in carbon steel. Liquid penetrant does the same job for surface-breaking flaws in austenitic stainless, aluminium and other non-magnetic materials, where MT will not work at all. Choosing PT for carbon steel because the crew has the kit is a downgrade in sensitivity; choosing MT for stainless is a scope that cannot find anything.

For the interior, the split is between radiography and ultrasound. Radiography images volumetric discontinuities such as porosity, slag, inclusions and incomplete penetration seen through the thickness, and produces a permanent image a non-specialist can look at. Ultrasound is more sensitive to planar flaws: lack of sidewall fusion, laminar tearing, and cracks oriented so that film shows nothing. It also measures through-wall height, which matters when the question is whether a flaw can be assessed rather than excavated. The RT vs UT weld inspection comparison works this through joint by joint.

Geometry decides more of this than material does. Single-sided access, thick sections, complex nozzle geometry, dissimilar metal joints and clad surfaces all constrain what can physically be applied. The method decision belongs in the inspection and test plan at drawing stage, agreed with the fabricator, rather than being made by whoever turns up on the day with whatever is in the van.

Who Sets Acceptance: AWS D1.1, ASME VIII and IX, API 1104

Acceptance criteria come from the fabrication code the contract invokes, and the three covering most US work are distinct in scope. AWS D1.1/D1.1M:2020 governs welded structural steel. Clause 8 covers inspection, Table 8.1 gives visual acceptance criteria, and Table 8.2 gives ultrasonic acceptance criteria, which rate an indication by comparing its amplitude to a calibrated reference level and measuring its length by the 6 dB drop technique. Annex H of the 2020 edition adds phased array from 3/16 in to 8 in, requires encoded scanning, and accepts no planar indication at any length.

ASME covers pressure equipment in two halves that buyers routinely merge. Section IX qualifies the welding procedure and the welder; it is about whether the weld can be made correctly. Section VIII Division 1 sets what examination is required and what is acceptable in the finished vessel, with UW-51 covering radiographic examination of butt welds and UW-51(a)(4) opening the ultrasonic route where the thinner material joined is at least 0.25 in. Section V supplies the examination methods themselves, and our ASME Section V reference sets out its article structure.

API Standard 1104 governs pipeline girth welds. Section 9 gives workmanship-based acceptance standards for the NDT methods, sized to the discontinuity type and length. Annex A provides an alternative route: fitness-for-purpose acceptance criteria derived from fracture mechanics, permitted only when the project specification, the ultrasonic procedure and the welding procedure together meet the prerequisites the annex sets out. Annex A is how a pipeline project accepts flaws that Section 9 would reject, and it is not something a contractor invokes unilaterally.

CWI vs NDT Level II, the Distinction Buyers Get Wrong

These are two different credentials, issued by two different bodies, proving two different things, and treating them as interchangeable is the most common qualification error in US weld inspection contracting. A Certified Welding Inspector is certified by the American Welding Society against AWS QC1. It is a visual welding inspection credential: the holder has demonstrated knowledge of welding processes, symbols, codes and visual acceptance criteria and passed a practical examination. AWS D1.1:2020 Clause 8.1.4 accepts a CWI, SCWI or CAWI, a CSA W178.2 equivalent, or an individual otherwise qualified to the Engineer's satisfaction.

An NDT Level II is not certified by a national body at all. Under ASNT SNT-TC-1A, which is a recommended practice rather than a standard, the employer certifies its own personnel against its own written practice, separately for each method. A UT Level II is qualified in ultrasonic testing; that certification says nothing about radiography, and a CWI card says nothing about either. ASNT and AWS do operate an arrangement allowing current CWIs and SCWIs to obtain visual testing certification under the ASNT Central Certification Program on the strength of the AWS examinations, but that route runs one way and covers VT only.

The practical boundary matters on site. A CWI verifies that welding conforms to an approved welding procedure specification. The CWI does not approve the WPS, does not make design or weld-size calculations, which belong to a registered professional engineer, and does not interpret radiographs or ultrasonic data on the strength of the CWI alone. Specify both credentials where both functions are needed, and ask for the certification records rather than accepting a statement that the crew is fully certified.

Shop vs Field: Two Different Inspection Problems

Shop inspection is the easier problem and the cheaper one. Fixed position, controlled environment, power, lighting, a calibration area, welds accessible from both sides, and an inspector who can watch fit-up, preheat and interpass temperature as the work happens rather than reconstructing it afterwards. Defect rates in a shop under continuous visual inspection are materially lower than in the field, and cost per weld inspected is a fraction of the field figure, because there is no travel, scaffold or standby time buried in it.

Field inspection changes every variable. Position welds instead of rolled, weather, restricted access, single-sided joints, surface condition, and a schedule where the inspector is one constraint among twenty. Radiography in the field brings exclusion zones and permits that stop other trades, which is why encoded ultrasound has taken so much field weld volume in US construction. The right response is to move as much of the inspection burden into the shop as the design allows, which is a fabrication decision as much as an inspection one and worth taking before the drawings are frozen.

It also changes what coverage means contractually. A 10 percent radiographic examination requirement is straightforward in a shop. In the field, unless the contract states which 10 percent, how it is selected, and what happens when a rejected weld triggers additional examination, the requirement is an argument waiting to happen. Write the selection rule and the escalation rule into the scope, not into the correspondence afterwards.

The RT-vs-UT Decision and Where PAUT Fits

The oldest live question in weld inspection is whether to shoot or scan. On thin sections with expected porosity and slag, radiography is direct, familiar, and produces an image that an owner, an engineer and an insurer can all look at. On thick sections, on planar defects, and anywhere flaw height matters, ultrasound outperforms it. Add the operational cost of a radiation exclusion zone on a live plant or a busy fabrication floor and the calculation usually resolves in favour of ultrasound well before the technical argument does.

Phased array sharpens that. Encoded PAUT sweeps a range of angles from a single probe position, images the weld volume, records everything against encoder coordinates, and sizes flaw height. For structural work, AWS D1.1:2020 Annex H is the acceptance route. For pressure equipment, ASME Section VIII Division 1 UW-51(a)(4) permits ultrasonic examination in lieu of radiography above 0.25 in, routing through Division 2 paragraph 7.5.5 to Section V Article 4 Mandatory Appendix VIII, which requires automated or semi-automated scanning with computer-based data acquisition. Manual phased array does not qualify for that route.

So the practical rule for a buyer is simple: any ultrasonic result intended to substitute for radiography must be encoded. Ask for the encoder setup, the scan increment and the raw data files in the scope and the ambiguity disappears before the first weld is scanned. Our phased array inspection services page covers the technique, the scan plan and the deliverable in detail.

What a Complete Weld Inspection Record Contains

A weld inspection record has to let someone who was not there reconstruct what was examined, by whom, against what, and with what result. That means weld identification tied to a drawing and a weld map, welder or operator identification with their qualification record, the WPS and PQR references, the method and procedure number and revision for each examination, the equipment and calibration record and calibration block, the extent of coverage actually achieved, the indications found with location and size and characterisation, the acceptance standard applied by clause number, the disposition, and the examiner's certification.

The two items most often missing are coverage and clause. Coverage, meaning what proportion of the weld volume the examination actually reached as opposed to what the procedure intended, is the difference between an examination and a gesture, and it is the first thing a competent reviewer checks. The acceptance clause matters because accepted against Table 8.1 and accepted against Table 8.2 are different statements about the same weld, and a record that does not say which one is unusable in a dispute.

Repair and re-examination history closes the loop. A weld that was rejected, excavated, rewelded and re-examined must carry that history, because the repaired region is often examined to a different extent than the original. Where records arrive from multiple subcontractors in different formats, they need normalising before anyone can audit them, which is what NDT report validation and the checklist in what makes an NDT report defensible exist to do.

How to Scope a Weld Inspection Contract

Name six things and the bids become comparable. One: the governing code and edition, whether AWS D1.1:2020, ASME Section VIII Division 1, or API 1104, with the acceptance table or section cited. Two: the weld population by joint type, material, thickness and count, on a marked-up drawing. Three: the method and extent per joint category, with the selection rule for partial coverage. Four: personnel qualification, separately for visual inspection and for each NDT method. Five: shop, field, or both. Six: the record format and who owns the raw data.

Then name the escalation rule. What happens when a weld is rejected: how many additional welds by the same welder are examined, how a repair is re-examined, and who pays for each. Codes address this in different ways and contracts frequently do not address it at all, which turns a routine rejection into a commercial negotiation in the middle of a schedule. Settling it in the purchase order costs one paragraph and saves an argument on every project that has a rejection, which is every project.

Atlantis covers US shop and field weld inspection with CWI visual inspectors and NDT Level II and III personnel across RT, UT, PAUT, MT and PT, delivers records in an auditable format, and reviews or validates weld inspection records produced by others. Where the requirement is capability rather than coverage, NDT training and ASNT Level III consulting build it in-house. Scope and pricing on request, starting at contact.

Request a consultation