Facing an Offshore NDT Audit: What the Level III Signs, and What the Auditor Opens First

What an offshore NDT audit is actually sampling

An auditor is not evaluating whether your inspectors are good people or competent in the abstract. They are testing whether a claim can be traced. The mechanism is almost always the same: pick a completed report and walk backwards. Which procedure produced it, at which revision? Was that procedure qualified, and by whom? Which examiner performed it, at what level, certified under which written practice, with what vision test on file? What equipment was used, and was it in calibration on that date? What acceptance standard was applied, and does the report name it?

If any link in that chain is missing, the finding is not against the individual report. It is against the system, because the same gap applies to every report produced under the same arrangement. That is why a single missing vision test can escalate into a corrective action covering a year of work, and why auditors deliberately sample a handful of records rather than attempting to review everything. The sample is a probe into the system, not an assessment of the sampled item.

Offshore adds a further dimension that onshore audits often skip. The auditor also wants to know that the technique was qualified for the condition it was actually used in. A UT procedure qualified on clean, dry, machined coupons does not on its own demonstrate capability on a splash zone member carrying six millimetres of coating and marine growth, examined from a rope access position in moving water. The evidence has to close that gap explicitly, in writing, before the audit rather than during it.

The evidence chain, link by link

The written practice is the root document and the one auditors read first, because every other claim derives its authority from it. Built on SNT-TC-1A, it defines the methods the employer certifies, the levels, the training and experience requirements for each level and method, the examinations and how they are graded, the vision requirements and their interval, the recertification period, and the conditions under which personnel transferring in from another employer may be certified.

Below that sit procedures and their qualification records, personnel certification files, equipment calibration records traceable to national standards, and the examination records with their raw data. Each layer must be internally consistent with the layer above. A written practice requiring forty hours of training for UT Level II and a certification file showing twenty four hours is a finding that needs no technical judgement whatsoever to raise, and it invites the auditor to look much harder at everything else in the file.

The vision requirement is the most sampled single item in the entire file, precisely because it is unambiguous, dated and instantly checkable. It also fails more often than anything else, usually because somebody left, somebody joined, and the tracking spreadsheet was never updated. Holding certification status, expiry dates and calibration due dates in a system with alerts rather than in a spreadsheet on a shared drive is one of the main reasons operators and contractors move this data into an inspection data management system.

BSEE, class and the operator — three doors into the same records

On the US Outer Continental Shelf, BSEE inspects under 30 CFR Part 250 and can request fabrication, welding and examination records supporting safety and pollution prevention equipment and structural integrity. Class societies audit against their own rules and survey regimes and are chiefly interested in whether the survey scope was met and whether the examination supporting it was performed by qualified people working to a qualified procedure. Operators run their own supplier audits against contractual, technical and HSE requirements that are usually stricter than either.

The three have different remits but they read the same files, and inconsistency between what you told each of them is itself a finding. A procedure revised to satisfy a class comment but never reflected in the written practice, or a personnel certification presented to an operator that does not match the record held by the employer, will surface the moment the second auditor arrives and asks for the same document from a different angle.

Verification and certification schemes add a further layer, where an independent certifying entity reviews the scheme covering safety-critical elements and their performance standards. The NDT contribution to that scheme is the evidence that a stated inspection was genuinely capable of finding what the performance standard says it must find, at the size and orientation that matters. That is a technique capability argument rather than a compliance statement, and it is one a Level III has to be able to make on paper and defend under questioning.

Splash zone and subsea: the records that are hardest to defend

The splash zone is the hardest region on an offshore structure to inspect and the easiest to write up badly. Coating condition varies across the tidal range, access is by rope or from a boat in moving water, surface preparation is limited by what can be achieved safely, and the same member may be dry, wet and submerged during a single shift. A wall thickness reading taken through an unknown coating thickness with a technique that assumes bare steel is a number with no defensible uncertainty attached to it, however carefully it was recorded.

Subsea inspection generates a different evidence problem. Data is produced by divers or ROVs, video is the primary record for close visual and general visual inspection, and cathodic protection readings, flooded member detection, ACFM and ultrasonics are taken by operators working in conditions that cannot be reproduced onshore. The audit questions are predictable: was the technician qualified for that technique and that deployment method, was the equipment calibrated for use at that depth and temperature, was reference calibration performed subsea or topside, and can the position of a reading be established from the video and survey record?

For structural weld examination, API RP 2X addresses ultrasonic and magnetic particle examination of offshore structural fabrication and, importantly, the qualification of the technicians who perform it. Underwater welding and its examination fall under AWS D3.6M. Where an operator wants inspection history located on the actual member rather than buried in a report folder, tying readings and anomalies to structural geometry through digital twins makes the coverage argument far easier to sustain across successive campaigns and successive contractors.

What the Level III signs, and what the Level III does not sign

The Level III's signature belongs on a defined set of documents: the written practice, NDT procedures and their qualification records, examination questions and grading, personnel certification within the employer's written practice where the practice designates the Level III to do so, technique capability statements, and independent reviews of recorded data. Each of these is a technical judgement about method capability or about a person's demonstrated competence, and each is defensible on evidence the Level III holds.

It does not belong on documents that carry somebody else's statutory or professional responsibility. Atlantis is not a PSM auditor and does not sign off a mechanical integrity programme as compliant. Atlantis is not the API inspector of record for a vessel, tank or piping circuit — that role belongs to a certified inspector engaged in that capacity, whose signature carries a different legal meaning entirely. And Atlantis does not deliver API 510, 570 or 653 inspector certification training; those are API ICP credentials with their own defined examination route.

Being explicit about that boundary is not a limitation, it is what makes the technical authority credible in the room. An auditor who sees a Level III signing things outside the Level III's remit immediately begins to doubt the signatures that were properly inside it. Where an operator needs the NDE side of an integrity programme to be technically defensible, the Level III supports it as technical authority alongside — never instead of — the inspector of record and the operator's own engineering authority.

Gap review before the audit versus remediation after a finding

A pre-audit gap review is voluntary, scoped and fast. It applies the same sampling logic an auditor would use, deliberately on a hostile basis, and produces a defect list with named owners and dates. Nothing raised in it appears on any external record. Missing training hours can be completed, expired vision tests renewed, procedures requalified and reports reissued with the acceptance standard properly named, all before anyone external opens the file and forms a view of the organisation.

Remediation after a finding is a different exercise with a materially different cost. The finding is on record, it usually carries a due date, and the corrective action must address root cause and demonstrate effectiveness rather than simply fix the sampled instance. If the finding is systemic — an unqualified procedure, an uncontrolled written practice, an examiner certified outside the practice — the scope of affected work has to be defined and often re-examined. A single vision test finding, closed properly, can require review of every report that examiner produced in the period.

The sequencing argument is straightforward: the work is the same work, but doing it early is a project and doing it late is an incident. Where a written practice needs rebuilding or procedures need qualification, that is Level III consulting and procedure development. Where examiner training hours or method coverage fall short, structured NDT training to SNT-TC-1A and ISO 9712 closes the gap against your own written practice rather than against a generic syllabus that will not match it.

Findings that recur offshore

The recurring set is remarkably stable across operators and contractors. Vision tests lapsed or missing for one or two examiners. A written practice naming methods the company no longer performs while omitting ones it does. Procedures issued at a revision that does not match the revision cited on the reports. Calibration blocks with no traceability certificate. Equipment calibrated but the certificate held by the hire company rather than in the file. Reports naming no acceptance standard. Diver and ROV records with no positional reference. Subcontracted NDE accepted with no review of the subcontractor's own practice or personnel.

Two findings deserve particular attention because they undermine everything else in the file. The first is a written practice copied from a template and never reconciled with what the organisation actually does; an auditor establishes that within ten minutes and then distrusts every document that claims authority from it. The second is a report whose conclusion is not supported by the raw data behind it, which is why an independent review of inspection reports and data is worth running on a sample well before an audit rather than after one.

Subcontracted inspection is the blind spot that grows fastest as campaigns scale. When a contractor subcontracts rope access ultrasonics or subsea NDE, the obligation to demonstrate competence does not transfer with the work. Somebody must have reviewed the subcontractor's written practice, procedures and personnel records, and must be able to produce that review on request. In most audit failures involving subcontractors, that review was either never carried out or was carried out informally and never written down.

How audit representation works on the day

Representation is narrow and specific by design. The Level III answers technical questions about method capability, procedure qualification, examination coverage, personnel competence and data interpretation. They do not answer for commercial terms, HSE performance, or management system elements outside the NDT scope, and they should not be introduced to the audit team as though they do. An overreaching representative invites questions the organisation is not prepared to answer.

The value is in answers that are consistent and evidenced. When an auditor asks why a particular technique was considered capable of finding a fatigue crack at a nodal weld toe beneath coating, the answer needs to reference the qualification record, the demonstration specimen and the detection threshold achieved — not the examiner's years of experience. An unsupported answer delivered confidently is worse than saying the evidence will be produced within the day, because the auditor then samples harder and wider.

Practically, preparation matters more than performance on the day. Files assembled by evidence chain rather than by folder structure, one sample walked end to end in advance to find the break before the auditor does, a named owner for each document type, and a short technical brief for everyone who will be in the room. Where an operator or contractor wants that preparation run independently ahead of a client, class or regulator audit, a scoped gap review timed against the audit date is the usual starting point.

Request a consultation