What an auditor actually does with your thickness register

An audit-ready CML and TML registry does more than store thickness. For any reading an auditor points at, it produces the technician's certification valid on that date, the procedure revision in force, the instrument calibration, the derived corrosion rate with its inputs, the next inspection due date, and the approval that set it. The evidentiary chain is the deliverable; the number is just its endpoint.

Petrochemical sites are audited from several directions at once. OSHA process safety management at 29 CFR 1910.119(j) requires inspection and testing of process equipment following recognised and generally accepted good engineering practice, on a documented schedule, with deficiencies corrected under 1910.119(j)(5). EPA's risk management programme at 40 CFR 68.73 repeats the obligation. A jurisdictional inspector may work to the National Board Inspection Code, your insurer runs its own survey, an ISO 9001 registrar samples the quality records, and a joint-venture partner or toll customer audits the asset they do not own. None of them read the whole register. Each picks a small sample and follows it backwards until something breaks. That is why an audit is not passed by having good data; it is passed by having data whose provenance survives being pulled on. A register that cannot answer "prove this number" in one traversal converts a two-day audit into a six-week finding.

Source: Written against 29 CFR 1910.119(j) and (l) (OSHA process safety management, mechanical integrity and management of change), 40 CFR 68.73 (EPA RMP mechanical integrity), API 510 and API 570 inspection intervals and owner-user inspection organisation requirements, API 571 (damage mechanisms), API 579-1/ASME FFS-1 (fitness-for-service), API RP 583 (corrosion under insulation), NACE/AMPP SP0170 (protection of austenitic stainless steels from polythionic acid stress corrosion cracking during shutdown), ASME PCC-2 (repair of pressure equipment), ASNT SNT-TC-1A and CP-189, and ISO 9712.

Technically reviewed by Anoop Rayavarapu — ASNT NDT Level III (UT, RT, MT, PT, VT, ET) · API 653 · ISO 9001:2015 Lead Auditor
The evidentiary chain an auditor follows backwards from a single thickness number
Audit stepQuestion askedRecord that must existWhere registers commonly break
Identify the locationWhich physical point produced this value?CML with drawing reference, component, position and a photographLocation described only by line number, so nobody can prove it is the same point
Find the examinationWhen was it taken and under what procedure?Reading with date, procedure revision, technique and extentProcedure cited by title only, with no revision, and the revision has since changed
Qualify the examinerWas the technician certified for this method on that date?Level II certification, method, expiry and annual vision recordCertification checked as valid today rather than valid on the examination date
Verify the instrumentWas the gauge calibrated and on what reference?Instrument serial, calibration status, calibration block or step wedge usedSerial number captured, calibration status never linked to the date of use
Reproduce the calculationHow was remaining life derived?Readings used, exclusions, retirement thickness in force, computation dateRate stored as a static value; inputs no longer recoverable
Close the loopWhat was decided and by whom?Next due date, approval signature, any deferral evaluation, deficiency recordDue date typed by hand into a planner, disconnected from the calculation that justified it
Auditors sample. A single unrecoverable link in one sampled chain typically widens the sample rather than ending the enquiry, which is how a narrow question becomes a programme-level finding.

Who actually audits a petrochemical thickness register

Petrochemical sites carry more audit exposure than almost any other process industry, because several bodies with different questions arrive on different cycles. OSHA process safety management at 29 CFR 1910.119(j) obliges the operator to inspect and test process equipment following recognised and generally accepted good engineering practice, on a documented schedule, using trained personnel, with results recorded. EPA's risk management programme at 40 CFR 68.73 restates the same duty for covered processes. Neither regulator prescribes software, but both effectively require a record that can be reconstructed on demand.

Alongside the regulators sit parties with commercial rather than statutory interest. A jurisdictional inspector may work to the National Board Inspection Code for in-service pressure equipment. The property insurer runs a survey whose findings drive premium and, occasionally, coverage conditions. An ISO 9001 registrar samples the quality records around inspection. And in a sector full of joint ventures, tolling arrangements and long-term offtake, the customer or partner who does not own the asset audits it anyway, often with more persistence than anyone else because their product depends on the plant running.

What unites them is method. None of them reads the register. Each one samples it and follows the sample backwards. The register is therefore not judged on how much it contains but on whether any given entry can be traced to its origin without a phone call, an archive box or a retired engineer's memory.

The ten-CML sample: how an audit really runs

A competent auditor opens the inspection plan, picks a unit, and asks for ten monitoring locations chosen by them rather than by you. For each one they walk a chain: show me the location and prove it is the physical point you say it is; show me the last examination and the procedure revision under which it was performed; show me the certification of the technician valid on that date; show me the instrument and its calibration status at the time of use; show me how remaining life was derived from these readings; show me the next due date and the approval that set it.

The chain has no weak-link tolerance. If certification is checked against today rather than the examination date, the auditor has found a systemic control gap and will expand the sample rather than close it. If the procedure is cited by title with no revision, and the procedure has since been revised twice, the examination cannot be shown to have been performed under a qualified method. If the corrosion rate is a stored value with no recoverable inputs, the due date resting on it is unsupported for every circuit in the plant, not just this one.

The design implication is straightforward and rarely implemented: the register must be traversable in the auditor's direction, from a number backwards to its evidence, not merely in the engineer's direction from an asset forwards to a report. A system that can produce a beautiful inspection dashboard but needs three departments to answer "prove this number" is built the wrong way round.

Petrochemical damage mechanisms that thickness alone will not catch

A register that models only general wall loss will pass its own internal checks and still leave the plant exposed, because much of what damages petrochemical equipment does not thin it uniformly. Chloride stress corrosion cracking of austenitic stainless under wet insulation produces no measurable thickness change until the moment of leak. API RP 583 places the corrosion-under-insulation risk band for carbon steel roughly between minus 12 and 175 degrees Celsius, which covers a large fraction of the insulated inventory on an olefins or aromatics unit and almost none of the equipment anyone thinks to prioritise.

Polythionic acid stress corrosion cracking is the shutdown-specific mechanism. Sensitised austenitic stainless exposed to sulfides forms polythionic acids when air and moisture reach it during a turnaround, and NACE/AMPP SP0170 sets out the neutralisation and purge practices that prevent it. The register's job is to hold that as a scheduled protective action tied to specific equipment, with evidence it was executed, because an auditor examining a shutdown will ask what was done and who confirmed it.

Caustic gouging and caustic stress corrosion cracking in caustic wash and treating sections, carburisation and metal dusting in syngas and cracking service, and high-temperature hydrogen attack where hydrogen partial pressure and temperature approach the API RP 941 curves all share the same property: the credible monitoring evidence is not a thickness value. A registry that cannot attach a replication result, a hardness survey, a permeability measurement or a wet fluorescent magnetic particle report to a monitoring location has quietly defined the plant's damage model as whatever a thickness gauge can see.

Creep, carburisation and the t-min that is not a pressure calculation

Nowhere is this clearer than an ethylene cracker radiant coil. The coil operates in the creep range by design, and its life is consumed by mechanisms that a pressure-based minimum thickness calculation does not describe. Diametral growth from creep, internal carburisation forming a brittle layer that changes the tube's magnetic response, and coke-driven tube metal temperature escalation together determine when the coil comes out. Retirement decisions rest on measured diameter increase, carburisation depth and, where taken, replication evidence.

A register built on the assumption that every monitoring location produces a thickness in inches or millimetres cannot hold any of this. In practice, plants respond by keeping the coil data somewhere else, which reproduces the original problem: two records, two owners, no single traversable chain when the auditor asks about the furnace. The correct structure is a monitoring location whose measurand is declared per location — wall thickness, outside diameter, hardness, carburised depth, ovality — each with its own units, limits and trending rules.

The same structural need appears elsewhere on the site. Tank settlement surveys, exchanger tube eddy current results expressed as percentage wall loss, flange face condition, and bolt tension checks are all periodic condition measurements attached to a physical location. Once the register accepts a general measurand rather than assuming thickness, the plant gets one interval engine and one approval workflow across all of them, which is precisely what makes a sampled audit traversable.

Intervals, due dates and the documented deferral

Under API 510 and API 570, an inspection interval is a calculated output constrained by a code maximum, commonly derived from remaining life with a fraction applied, and capped regardless of how favourable the arithmetic looks. The crucial property for audit purposes is that the due date is derived, not typed. When an auditor asks why a vessel is due in 2029, the acceptable answer names the readings, the corrosion rate, the remaining life and the rule applied. The unacceptable answer is that it was in the planner.

Deferrals are where most programmes actually fail. Turnarounds move, access is unavailable, a critical spare has an eighteen-month lead time, and the inspection slips. Slipping is not by itself a violation; slipping without a documented engineering evaluation is. A defensible deferral records the technical justification, the current corrosion rate and remaining life, any compensating measures such as tightened integrity operating windows or increased process surveillance, the approver and their authority, and the new date. It is an object in the register with its own approval, not an edited field.

Auditors have learned to look specifically for this pattern, because it is where schedule pressure and technical judgement collide. A register that allows a due date to be edited directly, with no deferral workflow and no record of who changed it, hands the auditor a finding that reaches beyond the individual equipment item to the credibility of the whole mechanical integrity programme.

Technician qualification is part of the record

Under ASNT SNT-TC-1A, certification is employer-based: the employer writes a written practice, and qualifies and certifies its own personnel against it. ASNT CP-189 and ISO 9712 offer alternative frameworks, and many petrochemical owners specify which applies in their contracts. In every case, the certification belongs to an organisation and a date, which means that a contract technician who examined your piping in 2023 was certified by their employer under their employer's written practice as it stood in 2023.

The register must therefore evaluate qualification against the examination date, not the present. This sounds obvious and is very commonly implemented backwards, because storing a link to the technician's current certificate is easier than snapshotting the certificate that was valid at the time. The failure is silent: the data looks perfect until an auditor samples a four-year-old reading and finds the certificate on file was issued last spring, which proves nothing about the person who took the measurement.

The same logic applies to the written practice itself, the procedure revision, and the annual vision acuity record that most written practices require. Holding all four as dated artefacts is not bureaucratic excess. It is the difference between being able to demonstrate that an examination was performed by a qualified person under a qualified method and merely asserting it.

Deficiency tracking and the loop back to management of change

The most-cited mechanical integrity gap is not a missing inspection; it is an inspection that found something and a record that does not show what happened next. OSHA at 1910.119(j)(5) requires that equipment deficiencies outside acceptable limits are corrected before further use, or in a safe and timely manner when necessary means are taken to assure safe operation. Auditors read that as a demand for a closed loop with dates on both ends.

In a register, that loop should be explicit. A reading below retirement thickness raises a deficiency attached to the monitoring location. If a fitness-for-service assessment under API 579-1/ASME FFS-1 is performed, its result and the operating envelope it supports attach to the deficiency. If a temporary repair is applied under ASME PCC-2, the repair type, its intended duration and the compensating measures attach as well. If operating limits are reduced pending permanent repair, that is a change to the process and belongs in a management of change record under 1910.119(l).

The audit question that catches most sites is the last one: when was the temporary measure removed? Clamps and reinforcing repairs installed as ninety-day expedients have a way of persisting for years because nothing in the record ever demanded their closure. A register that gives every temporary measure an expiry, escalates it when the date passes, and refuses to let the deficiency close without a documented permanent resolution turns that recurring finding into a routine report.

How to evaluate a register against an audit you have not had yet

Run the audit in the demonstration. Choose five monitoring locations from your own plant, give them to the vendor, and ask them to walk the full chain on screen for each: the physical location and its evidence, the examination and its procedure revision, the technician's certification as it stood on that date, the instrument and its calibration at time of use, the derivation of remaining life, the due date and the approval. Time it. If the chain takes longer than the auditor's patience, the software has not solved the problem you are buying it for.

Then test the awkward cases deliberately. Ask to see a deferral created and approved. Ask to see a reading corrected and observe whether the audit trail survives and the downstream calculations move. Ask what happens when a technician's certification expires between examination and report approval. Ask for a read-only auditor role that can traverse the record without touching it, because handing an auditor a full-privilege login is its own finding.

Finally, ask for the audit pack: a single export covering a selected set of locations with every supporting artefact attached, generated without a developer. Petrochemical operators typically face several audits a year from different parties asking overlapping questions, and the difference between a register that produces that pack in an afternoon and one that requires a two-week manual assembly is measured in engineering weeks. Atlantis builds this module on a fully customisable Odoo-based platform, and a demonstration against your own sampled locations can be arranged through info@atlantisndt.com.

What does an OSHA PSM auditor look for in thickness records?

Evidence that inspections and tests were performed on process equipment following recognised and generally accepted good engineering practice, on a documented schedule consistent with manufacturer recommendations and prior operating experience, by qualified personnel, with results documented. The auditor then checks 1910.119(j)(5): that equipment deficiencies outside acceptable limits were corrected before further use, or in a safe and timely manner with necessary interim measures documented. Records that show the finding but not the correction are the classic gap.

How long must thickness measurement records be retained?

For the life of the equipment, in practice. Inspection codes rely on comparison against original and prior measurements, so discarding an early reading destroys the long-term corrosion rate that justifies the current interval. Jurisdictional and insurer expectations vary, but no serious owner-user programme retains less than the full history plus the calculations derived from it. The register should make deletion structurally impossible and treat superseded readings as voided-with-reason rather than removed.

What makes an inspection interval deferral defensible?

A deferral is defensible when it is a documented engineering evaluation rather than a schedule slip. It needs the technical basis for extending, the corrosion rate and remaining life supporting it, any compensating measures such as increased process monitoring or an integrity operating window tightening, a named approver with the authority to approve it, and a revised due date. A deferral that exists only as a changed date in a planner is an audit finding waiting to be written.

Why does creep change how a petrochemical retirement thickness is set?

In an ethylene cracker radiant coil, wall loss is not the controlling degradation. Long-term exposure above the creep range produces diametral growth, creep voiding and carburisation of the inner surface, and the tube is retired on diameter increase, carburised layer depth or replication evidence long before a pressure-based minimum thickness is reached. A register that can only hold thickness will silently report a healthy remaining life on a coil that is at end of life.

Do technician certifications belong in the thickness record?

Yes, and validated against the examination date rather than today. Under ASNT SNT-TC-1A the employer certifies to its own written practice, so a contractor's technician is certified by that contractor, with method, level, examination results and annual vision acuity all part of the file. An auditor sampling a 2023 reading wants the certification that was current in 2023. Registers that store a live certificate reference silently lose that history at the next renewal.

How does a register connect a deficiency back to management of change?

By making the deficiency an object rather than a comment. A reading below the retirement thickness should open a deficiency linked to the CML, carrying the fitness-for-service assessment under API 579-1/ASME FFS-1 if one was performed, the repair or temporary measure under ASME PCC-2, the operating restriction imposed in the interim, the management of change record raised under 1910.119(l), and the date the interim measure was removed. Every one of those is a question an auditor asks in sequence.

Request a consultation

Built for any business that runs on operations

Most companies do not fail at their craft. They lose time, margin and goodwill in the gaps between the tools they use to run the place — a quoting spreadsheet that does not talk to the job sheet, a job sheet that does not reach accounts, and a compliance folder nobody can search when a client asks. Atlantis closes those gaps by putting the whole operation on one platform, so information is entered once and everything downstream stays in step.

What you can run on it

  • Sales and CRM — leads, quotes, follow-ups and the pipeline that tells you what next month looks like.
  • Projects and job costing — plan the work, track the hours and materials against it, and see the margin while the job is still live rather than at final account.
  • Field and service teams — dispatch, schedules, mobile capture that works with no signal, and sign-off from site.
  • Inventory and purchasing — stock, suppliers, reorder points and goods receipt, joined to the jobs that consume them.
  • People — records, qualifications and licences with renewal reminders, timesheets, leave and payroll.
  • Quality and documents — procedures and forms under revision control, with the audit trail an inspection or accreditation body actually asks for.
  • Accounts — invoicing, expenses, multi-currency and the reporting your accountant stops chasing you for.

Affordable, accessible, fully customizable — and we mean each word

Affordable because the whole suite is included rather than sold to you a module at a time, and because implementation is done by people who have run operations rather than by a chain of subcontractors. Accessible because it runs in a browser and on a phone, works for a small team on day one, and does not need a specialist on staff to keep it alive. Fully customizable because your process is the thing that makes you competitive — the software should bend to it, not the other way round.

Industries we configure for

Service businesses and contractors, manufacturing and fabrication, trading and distribution, laboratories and testing houses, engineering consultancies, construction and facilities, and asset owners across energy, marine, aerospace and infrastructure. Inspection and testing is where we started, and it remains the sector we go deepest in — but the platform underneath is general-purpose, and most of what it does has nothing to do with inspection at all.

What happens when you get in touch

A short conversation, not a sales sequence. We ask how the business runs today and where it hurts, show you the platform doing that work, and send a written quote shaped to your region, your team size and the scope you actually need. No obligation, nothing to install first, and no pressure to decide on the call. Reach out and tell us what you are trying to fix.

Related: business management platform · inspection management software · choosing the right category of software · modules · by industry · asset integrity platform. Book a free consultation.