Defending a petrochemical NDT programme in front of an auditor

What the auditor actually samples

Auditors do not read a programme from the top down. They pull a thread. Typically it starts with an issued inspection report on a piece of equipment in a covered process, and from there the questions run backwards: which procedure produced this number, was that procedure demonstrated for this material and thickness, who held the probe, what was that person certified for on that date, was the equipment calibrated, and who reviewed the result before it was acted upon.

That means the sample size is small and the depth is total. Five reports examined completely tell an auditor more about a programme than a hundred skimmed, and any single broken link discredits the rest. It also means the programme is judged on its worst traceable record rather than its best, which is a very different standard from the one internal reviews usually apply.

Petrochemical audits add a second dimension. Because the NDT data feeds inspection decisions on equipment in processes covered by process safety management, the same records carry regulatory weight as well as technical weight. An auditor from a certifying body, a client quality organisation, a class society or a regulator will use different vocabulary, but all of them are testing whether the number on the report can be defended by evidence that existed at the time.

The evidence chain, link by link

The chain has a fixed shape. A written practice defines what qualification means at this employer. Training and documented experience feed examinations, which are graded and retained. Vision examinations are current at the stated intervals. The Level III signs the certification decision. A written procedure lists essential variables and is demonstrated for the configuration in use. Equipment is calibrated and traceable, with interim verification through the shift. The examination is performed, reported and reviewed by someone with the authority to review it.

Each link must have held on the date the work was performed. This is where programmes that look healthy today fail retrospective sampling: the certificates on the wall are current, but the report the auditor pulled was produced eleven months ago by a technician whose vision record had lapsed, or against a procedure revision that did not yet cover the configuration examined. Nothing about the present state repairs that.

The practical consequence is that the chain has to be maintained as a live record rather than assembled for inspection. Certification expiry, vision currency, procedure revision status, equipment calibration and the mapping of technicians to methods and scopes are all structured data. Holding them as live records rather than in a folder of scanned certificates is what makes a retrospective query answerable in minutes instead of weeks, and it is what turns a sampled audit into a routine exercise.

The written practice: a document audited against itself

The employer written practice is the most under-maintained document in most NDT programmes, and the easiest place for an auditor to find a finding, because it is audited against itself. Whatever it says, the records must match. A practice that specifies a colour vision interval the records do not follow is worse than one that specifies nothing, because it establishes a requirement the programme then breaches.

Typical gaps are structural rather than exotic. Methods in use that the practice never mentions, because phased array or computed radiography arrived after the last revision. Levels described in general terms with no examination content behind them. Recertification rules that no process actually executes. Limited certification scopes granted informally with no basis in the document. A practice inherited from a previous employer and never adapted to how this organisation actually works.

Fixing this is a technical exercise, not an editing one. It requires deciding what qualification genuinely means for each method as it is used here, writing examinations that test that, and then aligning the records. Written practice development, examination content and the certification framework behind it are core ASNT Level III consulting work, and the value is measured by how well the document survives being read aloud next to the files.

Procedures, essential variables and the demonstration record

A procedure that lists essential variables and has never been demonstrated is an unqualified procedure with good formatting. Under Section V the essential variables define the boundary within which the procedure has been shown to work, and any change to one requires requalification. In practice that means material and product form, thickness range, geometry and configuration, transducer or source parameters, couplant, surface condition, scanning parameters, equipment and software, and acceptance criteria.

The recurring petrochemical finding is site adaptation. A weld configuration turns up that the scan plan does not cover, a technician adjusts the technique sensibly and correctly, the examination goes ahead, and nothing is written down. Six months later the auditor asks for the demonstration record covering that configuration and there is none. The technique may have been perfectly sound; the programme still cannot demonstrate it was.

The remedy is cheap if it is anticipated. A defined route for technique variations, a template technical justification, a Level III available to approve within the same shift, and a supplementary demonstration where the variation is material. Programmes that provide that route get documented variations. Programmes that make variation approval slow and painful get undocumented ones, and the auditor finds those instead.

Contractor NDT qualification at turnaround scale

A major petrochemical turnaround can put several NDT contractors and hundreds of technicians inside the fence in a single week, each arriving with their own written practice, their own procedures and their own certification files. The reports they produce will carry the site's name for the life of the equipment record. Verification at mobilisation is the only point at which that population can be controlled at reasonable cost.

Verification means more than collecting photocopies. It means checking certification against the practice in force for the scope, confirming method, level and currency, confirming vision records, sampling practically on representative specimens rather than trusting paper, and reviewing each contractor's procedures for genuine equivalence instead of assuming one UT procedure is interchangeable with another. Where a contractor works to its own written practice while its reports are filed in the owner's equipment records, the relationship between the two practices needs to be explicit rather than assumed.

The scale of this argues for standardisation upstream. Where an owner sets a common qualification baseline and communicates it into the contractor community before mobilisation, verification becomes a check rather than a negotiation. Method-level competence itself is built through structured NDT training and certification to SNT-TC-1A and ISO 9712, which is a separate activity from the site-specific qualification that a turnaround scope demands.

Findings that recur in this industry

The list is short and remarkably consistent. Certification files with external certificates but no employer certification decision. Vision records lapsed across periods when interpretation was performed. Procedures revised without requalification. Interim calibration verification failures with no re-examination back to the last valid check. Reports issued before review, with review signatures added afterwards. Limited certifications granted verbally. A Level III of record who has never visited the site and cannot describe how the work is actually performed.

Two more are specific to contractor-heavy operations. First, technique sheets missing from report packages, so the report states a result without stating how it was obtained. Second, thickness data recorded without the corresponding calibration verification, which makes the readings unusable for corrosion rate calculation even though they were technically correct when taken. Both are documentation failures with direct integrity consequences downstream.

The pattern behind all of them is the same. Each was a reasonable operational decision made under time pressure by a competent person, and each destroyed the evidence rather than the work. Independent inspection report validation applied on a sampling basis during a turnaround, rather than after it, catches these while there is still time to re-examine and reissue.

What the Level III signs, and what nobody should ask him to sign

The boundaries matter, both technically and commercially. The Level III signs the written practice, NDT procedures and their qualification, examination content and administration, personnel certification within the practice, technical justifications for technique variation, and independent review findings on data and reports. In an audit he speaks to those subjects with authority and can produce the evidence behind each of them.

He does not sign the things that belong to other roles. Atlantis is not a process safety management auditor and does not perform or certify PSM compliance audits. It is not the API inspector of record for pressure vessels, piping or tanks, and does not issue or sign those inspection decisions. It does not sell API 510, 570 or 653 inspector certification training, and no engagement should be scoped as though it did. Fitness-for-service assessment and integrity operating window decisions belong to the owner-user and to separately qualified specialists.

Stating those limits in writing at the start of an engagement is not a disclaimer, it is a control. Audits go badly when a role was assumed rather than defined, and the person in the room discovers mid-question that someone expected them to own a decision they never made. A clear scope tells the auditor exactly who to ask about what, and that clarity itself reads as programme maturity.

PSM records and the mechanical integrity file

In a covered process, inspections and tests on process equipment must be performed by trained personnel, follow recognised and generally accepted good engineering practice, be conducted at a frequency consistent with manufacturer recommendations and prior operating experience, and be documented. The record content required is specific: the date, the name of the person who performed it, the identification of the equipment, a description of what was performed, and the results. Auditors read that list literally.

NDT reports frequently fail on the equipment identification and the description of what was performed. A thickness survey referenced to a drawing that has since been superseded, or condition monitoring locations recorded as approximate positions rather than as fixed identified points, cannot be repeated meaningfully. The next survey then measures somewhere else, the calculated corrosion rate becomes fiction, and the inspection interval is being set from noise.

Deficiencies outside acceptable limits must be corrected before further use, or in a safe and timely manner with necessary means to ensure safe operation. That makes the deficiency-to-closure trail part of the audited record rather than a maintenance detail. Holding inspection results, condition monitoring locations, deficiencies and their closure in mechanical integrity software keeps that trail intact between turnarounds, which is exactly the period during which paper records decay.

Gap review before the audit, remediation after the finding

A pre-audit gap review runs the auditor's own method against the programme, deliberately and on a sample. It pulls reports at random, traces each one backwards through the evidence chain, and records where the chain breaks. It reviews the written practice against the records rather than against intent, checks certification files against the practice, and examines procedures against the configurations actually examined. The output is a prioritised list with an honest assessment of which gaps can be rebuilt and which cannot.

The distinction between rebuildable and non-rebuildable matters more than anything else in this work. A lapsed procedure demonstration can be performed now and the affected examinations reviewed or repeated. A missing vision record cannot be created retrospectively, and backdating any record is fraud, not remediation. Where evidence genuinely cannot be reconstructed, the defensible route is to re-examine the affected work and say plainly what happened, which auditors treat very differently from a file that has been tidied.

After a finding, the same technical work happens under adverse conditions. There is a corrective action plan with dates, a root cause analysis the auditor will test, effectiveness verification at a follow-up, and often a client or regulator watching the schedule. The engineering is identical. Everything around it is worse. Starting a gap review three to six months before a scheduled audit or a major turnaround is the difference, and to scope one against your programme, contact the team at info@atlantisndt.com.

Representation in the room

Audit support is partly preparation and partly presence. Technical questions during an audit are usually answered badly not because the programme is weak but because the person answering does not know the boundary of the question. Asked why a particular technique was used on a particular weld, an inspection supervisor may explain the operational reason rather than the technical basis, and the auditor hears an admission where a justification existed.

A Level III present for the technical sessions answers those questions in the correct frame: this is the procedure, these are its essential variables, this is the demonstration record for this configuration, this is the technical justification for the variation, and this is who reviewed the data. Where the honest answer is that a gap exists, saying so directly with the remediation already in progress is consistently better received than an improvised defence.

Preparation for that role is specific work. It means knowing which reports the auditor is likely to pull, having traced them already, knowing where the programme is thin, and having decided in advance who answers what. Programmes that prepare this way tend to close audits with observations rather than findings, not because their evidence is perfect, but because they know precisely what their evidence says.

Request a consultation