The welding inspector's role (CSWIP/AWS CWI)

The most visible guarantee of fabrication quality is the weld itself. Behind that weld stands a name that usually stays in the shadows: the welding inspector. The service life of a pressure vessel, a bridge girder or a petrochemical pipeline often comes down to one question the inspector asked at the right moment, or one record they refused to sign. A welding inspector's job is not to look at a bead and call it pretty or ugly. It is a systematic engineering discipline that runs from pre-weld preparation to final acceptance, framed by standards, captured in records, and carrying legal responsibility. This article walks through the inspector's real role in the field, the limits of their authority, and the ethical ground they stand on, within the CSWIP 3.0/3.1/3.2 and AWS CWI framework.

1. Where the inspector stands: the independent eye on the production line

The welding inspector is not part of production; the inspector verifies production. The welder lays the bead and the foreman pushes the job forward. The inspector independently confirms that the work complies with the contract, the code and the procedures. That distinction matters: even under production pressure, an acceptance decision rests on the applicable acceptance criteria, never on the delivery schedule.

In practice the inspector works across three time windows: before welding, during welding and after welding. Experienced inspectors often say that most of the job is finished before the arc is struck, and they are right. A deviation caught before the arc is struck costs almost nothing to correct. Found afterwards, the same deviation means grinding, repair welding, repeat nondestructive testing (NDT) and lost schedule. That is why a seasoned inspector spends most of their effort on the pre-weld stage.

The inspector's authority to intervene in those three windows does not come out of thin air. It comes from the ITP (Inspection and Test Plan). The ITP fixes, line by line, which party (manufacturer, client, third party, notified body) plays which role at which fabrication step. Two types of point decide that authority. A hold point stops production until it is released; moving to the next step without release is a non-conformity in itself. A witness point requires notification, but if the party does not attend, production may continue after a defined waiting period. The sentence "work is stopped" only carries weight in the field if a hold point in the ITP backs it up. That is why an inspector's first day starts with the ITP, not with the weld.

One terminology note applies to this entire article. A discontinuity is any interruption in the homogeneous structure of the material or weld; the word is neutral and carries no judgment. An imperfection is a discontinuity that has been identified and measured (that is exactly what the naming and numbering in ISO 6520-1 is for). A defect is an imperfection that exceeds the acceptance criteria and therefore requires rejection and corrective action. An inspector never uses these three words interchangeably, because the difference between "there is a discontinuity" and "there is a defect" is a repair decision and a cost. Every defect is a discontinuity; not every discontinuity is a defect — the acceptance criteria decide which is which.

2. Pre-weld checks: the cheapest inspection happens here

The pre-weld stage is where the inspector adds the most value. These are the main items to verify:

WPS / pWPS verification: Is the approved Welding Procedure Specification (WPS) available on site, and is it supported by the relevant qualification record? That supporting record is called a PQR (Procedure Qualification Record) in the ASME system and a WPQR (Welding Procedure Qualification Record) in the ISO system. A pWPS (preliminary WPS) is a draft specification prepared for the qualification test and not yet qualified; never weld production joints to a pWPS. Does the WPS actually cover the work in hand — material group, thickness range, welding position and heat input? Procedure qualification is framed by ASME BPVC Section IX or ISO 15614-1, and the format of the specification by ISO 15609-1.
Welder qualification certificates: Does the welder hold a valid certificate for the position, process and material of this job? What matters is not the color or the date on the certificate but the range of qualification. In ISO 9606-1 the governing concept is the range of qualification; in ASME Section IX it is the QW-350 essential variables. Compare the coverage items one by one: welding process, filler material group (FM), material group, product type (plate/pipe), size (thickness t and diameter D), welding position, and weld detail (butt/fillet, single/double side, with or without backing).
How qualification stays valid: "Has it expired?" is not enough on its own. Under ISO 9606-1, validity depends on a confirmation of continued working signed every 6 months by the employer or responsible person. On the ASME Section IX side, records must show that the welder has welded with the process within the last 6 months. On site the certificate can look valid while that confirmation signature is missing; the chain of continuity is what the inspector really examines. For extension and re-test options, always work from the current edition.
Material conformity and traceability: Are the base material and the filler material traceable through a material certificate (EN 10204 3.1/3.2)? Is the batch/lot number of the filler material recorded and matched to the certificate? If the job requires alloy verification, has PMI (Positive Material Identification) been performed on site with an instrument (XRF/OES) to confirm the alloy composition?
Consumable conditioning: Have the filler materials been stored dry? For basic-coated electrodes and submerged-arc fluxes, does the re-drying temperature and time match the manufacturer's data sheet and the WPS (a typical order of magnitude for basic-coated electrodes is 300–350 °C for 2 hours; the exact value always follows the manufacturer's data)? Are the quiver temperature (typically in the 70–150 °C band) and the electrode's exposure time to atmosphere recorded? Oven and quiver records are the documentary side of your defense against hydrogen-induced cracking.
Fit-up: Are the groove angle, root gap, root face and internal misalignment (hi-lo) within tolerance? Is the joint face free of oil, corrosion product, paint and moisture? Were the tack welds made by a welder within the range of qualification, are they free of cracks, and were their ends properly ground before being welded into the joint? On materials that require preheat, tack welding is subject to preheat as well — this is exactly the item most often skipped in the field. A short tack, low heat input and rapid cooling together create ideal conditions for a tack crack.

At this stage the inspector's answer has to be a clear "acceptable" or "not acceptable". There is no "it will probably be fine".

3. Surveillance during welding: tracking the parameters live

Once the arc is struck, the inspector's role turns to monitoring and verification. The focus is on confirming that the variables defined in the WPS are actually being applied on the shop floor:

Welding parameters and heat input: Are current (A), arc voltage (V) and travel speed being measured, and is the resulting heat input within the WPS range? This is the calculation the inspector performs on site:

Q = k × (U × I × 60) / (v × 1000) — Q: heat input [kJ/mm], U: arc voltage [V], I: welding current [A], v: travel speed [mm/min], k: thermal efficiency factor.

Typical k values used in the logic of ISO/TR 18491 and EN 1011-1: SAW 1.0; SMAW, GMAW/MIG-MAG and FCAW 0.8; GTAW/TIG 0.6. An inspector who does not run the numbers can say "I checked the parameters", but cannot say "heat input is within the WPS".
The metallurgical consequence of heat input: High heat input coarsens the grain, lowers toughness and widens the HAZ. Low heat input increases the cooling rate, raising HAZ hardness and the risk of hydrogen-induced cold cracking; insufficient energy also sets the stage for lack of fusion. One caveat: lack of fusion is primarily a matter of technique, manipulation and access to the joint. Heat input makes it easier to occur, but does not determine it on its own.
Preheat: Has the minimum preheat temperature required by material group, component thickness and carbon equivalent been reached? The real argument here is where and when the temperature was measured. ISO 13916 defines the measuring distance and timing for preheat, interpass and post-heating temperatures; measurement is preferably taken on the face opposite the heated surface. Where that is not possible, allow a thickness-dependent soak time after the heat source is withdrawn, then measure at approximately 4 × thickness, minimum 75 mm from the groove (AWS D1.1 practice). A number ends the argument on site; "we measured from a suitable distance" does not.
Cold cracking is prevented by four factors: Preheat alone is no magic bullet. Hydrogen-induced cold cracking is prevented by controlling diffusible hydrogen content, tensile stress level, hardenable microstructure (carbon equivalent) and the temperature regime together. Low-hydrogen consumable management (oven and quiver discipline) and heat input control weigh as much as preheat.
Interpass temperature: Exceeding the upper limit creates different risks depending on the material. In austenitic grades with standard carbon content (304/316) the main risk is sensitization and the resulting loss of intergranular corrosion resistance. In low-carbon (304L/316L) and stabilized (321/347) grades that risk is largely designed out, so the dominant risks become hot cracking, grain coarsening and distortion. In duplex stainless steels, both the heat input window and the interpass upper limit (typically ≤150 °C) are the most critical parameters; exceeding them directly affects the ferrite/austenite balance and secondary phase precipitation. In low-alloy and quenched and tempered (QT) steels the main risk is loss of toughness. Knowing what is at stake in each material is what lets the inspector defend the correct limit.
Run sequence and cleaning: Do the root, fill and cap passes follow the sequence in the WPS? Is slag fully removed between passes? Whether the root run needs back purging depends on the material: carbon steel generally does not require it, while stainless steels, duplex grades, nickel alloys and titanium do. On those jobs the inspector should verify with an oxygen analyzer that residual oxygen has dropped below the target value, and on titanium should additionally check the color acceptance criteria for the root and cap surfaces and, where required, the use of a trailing shield.

At this stage the inspector does not intervene in production; the inspector documents the deviation immediately. When a deviation is seen and the ITP provides for it, work is stopped, the welding engineer is notified, and the corrective action goes on record.

The non-conformity and repair chain (NCR): A verbal warning is not a field record. The mandatory chain is this: raise and number a non-conformity report (NCR); define the root cause and corrective action; carry out the repair to an approved repair WPS (many codes require the repair procedure to be separately qualified); mark the repaired area on the repair map; re-examine the repaired weld using the same method and the same coverage as the original inspection; and respect the code limit on how many repairs are allowed in the same area. The NCR is closed with the signature of the verifying inspector.

4. After welding: visual testing is the foundation of all NDT

Visual testing (VT) is the foundation of every NDT method, and in practice it catches the majority of discontinuities. VT is cheap, fast and, done properly, extremely effective. Visual testing of fusion-welded joints is carried out to ISO 17637, which defines how the method is applied — pre-inspection conditions, lighting, access geometry and the inspector's vision requirements.

In post-weld VT the inspector looks for the following (using ISO 6520-1 nomenclature): surface cracks, surface-breaking porosity, undercut, overlap, excess weld metal, excessive convexity, root concavity (suck-back), burn-through, craters and crater cracks, spatter, stray arc (arc strike), and dimensional deviations (weld height, leg length, concavity/convexity, misalignment, angular distortion).

One boundary must be drawn clearly here: lack of root penetration can only be assessed by VT if the root surface is accessible — that is, if it can be seen from inside the pipe or from the back face on a single-sided weld. On a butt weld with no such access, root penetration is a question for volumetric methods such as RT or UT, not for VT. Stretching VT's capability beyond that condition produces an acceptance decision you cannot defend in the field.

PWHT and hardness verification. On pressure vessel and pipeline work, post-weld heat treatment (PWHT, stress relief) is one of the inspector's most demanding hold points, and it is a records job as serious as VT or NDT. Verify the following: thermocouple quantity and placement (suited to component geometry and code requirement); heating and cooling rates staying within the thickness-dependent limits; soak temperature and soak time complying with the code (typically along the lines of ASME BPVC Section VIII Div. 1 UCS-56 or ASME B31.3 Table 331.1.1); furnace or local heating band width and insulation; and correct matching of the cycle chart to the right job and the right weld. The chart enters the report annex as a signed record, not as a pencil note, and performing that chart review is the inspector's responsibility. Where required — for example on wet H₂S service covered by NACE MR0175 / ISO 15156 — hardness measurement after PWHT and comparison against the code or specification limit is part of the same step. Remember that post-heating and PWHT are different operations: post-heating is for hydrogen diffusion and does not replace stress relief.

Acceptance criteria are not a matter of opinion; they come from the applicable code. Classification and designation of weld imperfections is given in ISO 6520-1, and quality levels in ISO 5817. The scope of ISO 5817 is worth knowing as well: it applies to fusion welding of steel, nickel, titanium and their alloys, excluding beam welding, for t ≥ 0.5 mm. Quality levels for laser and electron beam welded joints are in the ISO 13919 series. Structural steel codes such as AWS D1.1 give their own visual acceptance tables. The inspector's job is not to say "looks good to me" but "accepted to ISO 5817 quality level B". An acceptance decision always rests on a written criterion.

5. Calling NDT and accepting the weld: know the method, know its limits

Visual testing tells you about the surface; NDT methods tell you about volumetric integrity. For welds, the umbrella standard for method selection, extent of testing and matching acceptance levels to quality levels is ISO 17635. That is what backs the phrase "call the right method in the right place", and where a contract states only a quality level, the conversion must be established through this standard.

The inspector's job is usually not to perform the NDT. The order of work is: verify that the method and extent comply with the ITP and the code; review the NDT procedure and the operator's certification scope; compare the NDT report against the code acceptance criteria; and tie the result into the fabrication record. There is a scope limit here: a CSWIP 3.1 or AWS CWI certificate does not, by itself, authorize interpretation of radiographic film or evaluation of UT/PAUT data. Interpreting film or data additionally requires ISO 9712 Level II in the relevant method (or, on the ASME/AWS side, Level II under the employer's written practice to ASNT SNT-TC-1A / CP-189).

PT (liquid penetrant testing): For surface-breaking discontinuities (application ISO 3452-1, acceptance ISO 23277).
MT (magnetic particle testing): For surface-breaking and near-surface discontinuities in ferromagnetic material (to roughly 1–2 mm depth, with sensitivity falling off rapidly with depth); application ISO 17638, acceptance ISO 23278. MT is not a volumetric method; relying on MT for a discontinuity at depth is a method error.
RT (radiographic testing): For volumetric internal discontinuities, with a permanent film or image record (application ISO 17636-1 for film and ISO 17636-2 for digital detectors; acceptance ISO 10675-1 for steel/nickel/titanium and ISO 10675-2 for aluminium).
UT (conventional, pulse-echo): For volumetric and planar discontinuities (application ISO 17640, acceptance ISO 11666). ISO 11666 applies to full penetration welds in ferritic steel and typically for t ≥ 8 mm; its acceptance levels are used in combination with the testing levels (A/B/C) of ISO 17640.
PAUT (phased array ultrasonic testing): Application ISO 13588, acceptance ISO 19285 (ferritic steel, full penetration joints, typically t ≥ 6 mm).
TOFD: Application ISO 10863, acceptance ISO 15626 (ferritic steel, typically from 6 mm upwards). Writing PAUT and TOFD on the same standard line as conventional UT produces the wrong procedure and the wrong acceptance level reference in the field. Accepting thin-wall or austenitic material against these standards is equally indefensible — those cases require special techniques and special acceptance criteria.

Matching the quality level to the acceptance level is the link most often skipped on site. Quality levels B/C/D of ISO 5817 correspond to acceptance levels 1/2/3 in the NDT acceptance standards (ISO 11666, 23277, 23278, 10675-1, 19285, 15626). Where the contract gives only a quality level, the inspector must make that conversion through ISO 17635 and fix it in writing. A signature saying "accepted to level B" is valid only once that mapping is on record.

Timing of inspection (delayed NDT). When you call NDT, the method is not the only criterion — time is too. In high-strength, thick-section and quenched and tempered (QT) steels, hydrogen-induced delayed cracking can appear hours after welding is finished, which is why codes impose a waiting period before inspection. AWS D1.1 applies a waiting period based on steel group and requires inspection of QT steels such as ASTM A514/A517/A709 Gr. 100 to start at least 48 hours after welding is complete. Many codes and project specifications apply 24 hours for thick-section low-alloy steels, and ASME B31.3 sets a similar hold for certain alloy groups (for example above 1¼Cr). The number of hours always comes from the relevant clause of the governing code; never quote it from memory. An NDT run before the waiting period expires and reported "clean" will not save a weld that cracks afterwards.

The inspector also has to know which method sees which discontinuity. RT is strong on volumetric discontinuities; on planar cracks it is reliable only when the crack is oriented close to the beam direction. A crack lying perpendicular to the beam, or lying flat, can easily be missed on a radiograph, whereas the same crack has a markedly higher probability of detection with ultrasonics — provided the scanning angle, probe configuration and scanning surface are chosen correctly. UT is orientation-dependent too, and a wrong angle, poor surface condition or insufficient scan coverage can miss that same crack. Confirm the extent of testing (spot or 100%), the acceptance level and the validity of the report against the code and the contract. The NDT operator must also be certified in the relevant method and level (usually ISO 9712 Level II/III, or Level II under ASNT SNT-TC-1A / CP-189 on the ASME/AWS side).

6. CSWIP 3.0/3.1/3.2 and AWS CWI: certificates and scope limits

Two major schemes document a welding inspector's competence:

CSWIP (TWI, UK origin; operated under ISO/IEC 17024 accreditation): The scheme's basic visual level is CSWIP 3.0 Visual Welding Inspector (Level 1). CSWIP 3.1 Welding Inspector (Level 2) documents standard welding inspection competence. CSWIP 3.2 Senior Welding Inspector (Level 3) is the higher level, meaning extended competence for evaluating procedures and records, preparing inspection plans, and supervising more complex work. Holding a valid 3.1 certificate is a prerequisite for sitting the 3.2 examination.
AWS CWI (American Welding Society): The Certified Welding Inspector documents inspection competence around AWS D1.1 and related codes; above it sits SCWI (Senior CWI) and below it CAWI (Certified Associate Welding Inspector). The scheme rests on two documents: AWS B5.1, which gives the qualification criteria for inspectors, and AWS QC1, which gives the certification rules. CWI candidates must also provide a documented eye examination.

Validity and renewal. An inspector who questions a welder's certificate validity while ignoring the status of their own is being indefensibly inconsistent. CSWIP certificates are valid for 5 years; at the end of year 5 they are renewed without examination, provided the employer documents work in the relevant field for at least four of the previous five years, and at year 10 a recertification examination is required. On the AWS side, the CWI certificate is renewed every 3 years (with continuing education / PDH records and an eye examination), and recertification is required at year 9. For the exact conditions, always work from the current edition of the relevant scheme document.

The critical point here is the scope limit. A certificate does not automatically give the inspector authority to approve a WPS, interpret NDT data or make welding engineering decisions. The inspector declares conformity — of the procedure, the welder, the material and the weld. Designing the procedure itself, approving the PQR/WPQR, selecting materials, setting the heat treatment regime and defining the repair strategy mostly belong to the welding engineer (IWE/EWE or the responsible engineer). A good inspector knows exactly where their own authority ends and when to hand the question to the engineer.

7. VT equipment and personnel requirements: what belongs in the bag

Visual testing is only as serious as the equipment behind it. A welding inspector's typical VT kit includes:

Weld inspection gauges: Used for excess weld metal and cap height on butt welds, leg length and throat thickness on fillet welds, and for undercut depth, misalignment and root gap measurement. These are separate tools: a bridge cam gauge and a multi-purpose Cambridge gauge have different designs and different measurement sets; they are not synonyms. Note too that most gauges do not measure fillet throat thickness directly; they derive it from leg length and convexity. Preserve that distinction when writing the report.
Lighting and a light meter: Under ISO 17637, the inspection surface requires a minimum of 350 lux of white light, with ≥500 lux recommended; enhanced visual testing calls for at least 1,000 lux. If you are working under ASME codes, ASME BPVC Section V Article 9 requires a minimum of 100 fc (1,000 lux) for direct visual examination. You cannot defend a requirement you cannot measure: carry a lux meter, and keep its calibration valid.
Access geometry: ISO 17637 requires the inspector's eye to be no more than 600 mm from the surface under inspection and to view the surface at an angle of at least 30°. If you must work outside those conditions, justify it and record it.
Vision test record: The precondition for a valid VT decision is the inspector's own eyesight. ISO 17637 requires near vision acuity sufficient to read Jaeger J1 / Times Roman N4.5 at a distance of not less than 300 mm, verified at least once a year; color differentiation is also checked separately. A current copy of that record belongs in the inspection file.
Magnifier: Usually 2×–5×, for examining fine cracks and porosity.
Temperature measuring tools: Contact thermometer, temperature crayon or non-contact measurement, for preheat and interpass temperature control.
Aids: Mirror (for inaccessible root or back faces), steel rule, caliper, flashlight, groove angle gauge, oxygen analyzer (on jobs requiring back purging) and dimensional check tools.

Calibration validity of the measuring instruments is also the inspector's responsibility. An acceptance decision given with an uncalibrated gauge or an unverified lux meter cannot be defended.

8. Reporting: no record means no inspection

There is a rule that holds in the field: an inspection that is not documented did not happen. The records the inspector produces are the legal and technical trace of the work. A good inspection report contains: job/project identity; joint number and location (isometric drawing / weld map reference); the WPS number applied; welder identity and qualification certificate number; material and filler material data (with batch number); inspection date and, where applicable, the waiting period observed after completion of welding; the method applied and the acceptance criteria (for example ISO 5817 quality level B and the corresponding acceptance level); the relevant NDT report numbers; the findings observed; the accept/reject decision; in case of rejection, the NCR number and repair map reference; the record of re-inspection after repair; the ITP line / hold point reference; and the inspector's name, certificate number and signature.

These links form the traceability chain; if one is missing, the chain breaks. Read welding symbols and weld representations to ISO 2553 / AWS A2.4, and frame the manufacturer's welding quality capability with the quality requirements of ISO 3834. Consistent, traceable, auditable records are the inspector's strongest defense at the next audit, in a warranty claim, or in a dispute.

9. Ethics and independence: the weight of a signature

The hardest part of inspection work is not technical but ethical. The inspector works on the same site and under the same pressure as the production team, yet an acceptance decision can never rest on a relationship, a schedule or a request to "let it go this once". Independence is the inspector's reason for existing.

The core ethical principles are clear: report only the actual condition; never accept an inspection you did not witness; stay inside your scope of authority; declare conflicts of interest; and never falsify records. An inspector's signature represents the competence and the honesty of the person behind it. When a weld that was accepted contrary to fact cracks in service years later, all that remains is that signature and the record beneath it.

From the field

Years ago, on a pressure vessel job, every welder certificate in the file looked valid. The fit-up was clean, the WPS was the right one, even the root run was flawless. Then we read the range of qualification line by line: the thickness range the welder held did not cover the shell thickness we were welding. The paper was green, but the range was wrong. The same file was also missing the last continued-working confirmation, the one the employer has to sign every six months. We stopped the work, re-tested the welder and lost three days. Production made its feelings clear. Months later that vessel passed its hydrostatic test on the first attempt. The lesson: keep your eye on the range of qualification and the continuity signature, not on the color of the certificate. An easy acceptance is rarely the cheap option, and the most expensive mistake is always the one you find after the arc is struck.

Related standards

Procedure and personnel qualification

ASME BPVC Section IX — Qualification of welding, brazing and fusing procedures (WPS/PQR) and welder/operator performance qualification (WPQ).
ISO 15614-1 — Qualification of welding procedures by test (steel/nickel, WPQR).
ISO 15609-1 — Welding procedure specification (WPS) format.
ISO 9606-1 — Qualification testing of welders (steel).
ISO 3834 — Quality requirements for fusion welding.
ISO 9712 — Certification of NDT personnel (ASME/AWS equivalent: ASNT SNT-TC-1A / CP-189).
AWS B5.1 / AWS QC1 — Welding inspector qualification criteria and the CWI certification scheme.

Acceptance criteria and imperfection classification

ISO 6520-1 — Classification and designation of imperfections in welds.
ISO 5817 — Quality levels for fusion-welded joints (steel, nickel, titanium; beam welding excluded, t ≥ 0.5 mm).
ISO 13919 — Quality levels for laser and electron beam welded joints.
AWS D1.1 — Structural welding code — steel (with its own visual and NDT acceptance tables and delayed inspection requirements).
ASME BPVC Section VIII Div. 1 / ASME B31.3 — Pressure vessels and process piping; PWHT and examination requirements.

NDT: umbrella, application and acceptance

ISO 17635 — NDT of welds: method selection, extent and acceptance level matching (umbrella standard).
ISO 17637 — Visual testing of fusion-welded joints (VT).
ASME BPVC Section V — NDE method requirements (Article 9: direct visual examination and lighting).
ISO 3452-1 / ISO 23277 — PT application / PT acceptance levels.
ISO 17638 / ISO 23278 — MT application / MT acceptance levels.
ISO 17636-1, -2 / ISO 10675-1, -2 — RT application (film / digital detector) / RT acceptance levels (steel-nickel-titanium / aluminium).
ISO 17640 / ISO 11666 — Conventional UT application and testing levels / UT acceptance levels.
ISO 13588 / ISO 19285 — PAUT application / PAUT acceptance levels.
ISO 10863 / ISO 15626 — TOFD application / TOFD acceptance levels.

Supporting standards

ISO 13916 — Measurement of preheat, interpass and post-heating temperatures.
ISO/TR 18491 and EN 1011-1 — Heat input calculation and thermal efficiency factors.
EN 10204 — Inspection documents for metallic products (3.1 / 3.2).
ISO 2553 / AWS A2.4 — Welding symbols and representation.
NACE MR0175 / ISO 15156 — Sour service material and hardness requirements.

Take it to the field: To keep the welding standards, procedure logic and field check steps in this article in your pocket — completely offline and free — take a look at the Doawise Welding Eng Guide app.

This article is for educational purposes and does not replace the official standard or the formal acceptance criteria. Always work from the current edition of the applicable code or standard and from the contract requirements.

At DoaWise we carry out welding engineering and inspection services to international standards — ASME, ISO and AWS above all — covering every step from pre-weld preparation to final acceptance and producing traceable, recordable and auditable results, so that the safety of both the manufacturer and the end user is backed by technical documentation.