Audit-ready thickness records for HRSGs and power piping
A CML and TML registry for gas and combined-cycle plants binds every reading to the technician, instrument, calibration check and procedure revision valid on the examination date, and keeps that binding immutable. It handles condition monitoring locations that are not thickness at all, such as hardness and replication points on Grade 91, and it produces the evidence chain an auditor asks for without re-assembly.
Power plant audits rarely dispute a thickness number. They dispute whether it can be traced. The finding reads as an inability to demonstrate that the examination was performed under a controlled procedure by a person qualified at the time, using equipment with valid calibration, and that the interval which followed was derived from that reading rather than carried forward. In a combined-cycle plant the difficulty is scale and cycling. A heat recovery steam generator holds thousands of tubes across LP, IP and HP circuits, harps, headers, drums and links, and the dominant thinning mechanism is flow-accelerated corrosion, which is local, geometry-driven and highly sensitive to chromium content, pH and temperature. Two-shifting adds thermal fatigue at attemperators, drum nozzles and header ligaments, so damage accumulates per start rather than per year. A registry that stores only a number per location cannot support any of that. It has to store the circumstances of the reading and the basis of the decision that followed it.
Source: Written against ASME Section I for power boilers, ASME B31.1 for power piping including its covered piping systems requirements, ASME Section IX for welding qualification, ASME PCC-3 for risk-based inspection planning, the National Board Inspection Code NB-23 for repairs and alterations, ASNT SNT-TC-1A and ISO 9712 for personnel qualification, material specifications SA-106 and SA-213 for wall tolerance, and EPRI guidance on flow-accelerated corrosion in fossil and combined-cycle plant.
| Audit question | Evidence expected | Registry field that answers it |
|---|---|---|
| Who took this reading, and were they qualified on that date? | Certification to SNT-TC-1A or ISO 9712 showing method, level, issue and expiry, plus current annual vision test, under the employer's written practice | Technician identity bound to the reading, with certification validity evaluated at the moment of capture rather than at audit |
| Was the instrument calibrated, and was calibration verified that shift? | Instrument make, model and serial, calibration certificate in date, the calibration block used, and pre-job and post-job checks | Equipment record attached to the reading set, carrying block identity and both check results with times |
| Which revision of the procedure governed this examination? | The controlled procedure as issued on the examination date, not the revision currently in force | Procedure number and revision stamped onto the reading, with the full revision history retained and readable |
| What was the previous reading, and is the difference real? | The prior value and date from the same grid square, the datum used, and the measurement uncertainty | Reading history held per grid square, with grid origin and datum carried forward between campaigns |
| How was the next inspection interval set? | The calculation, its inputs, the minimum thickness basis, and the named person who approved the result | Stored calculation inputs and an approval record, regenerated on demand rather than retyped into a report |
| Has any recorded number been changed since? | An amendment trail showing original value, replacement value, reason and approver | Append-only readings where amendments are new records, so overwriting is not possible in the first place |
Audits fail on traceability, not on thickness
Sit through enough inspection audits at generating stations and the pattern becomes predictable. The auditor does not challenge whether the wall is 6.1 millimetres. They pick one reading out of several thousand and walk it backwards. Which procedure was in force that day, and can you show the revision as issued rather than the revision on the shelf now? Who held the probe, and was their Level II current on that date, including the annual vision test the written practice requires? What instrument, what calibration block, what pre-job and post-job checks? Then the same walk forward: what calculation used that reading, what minimum thickness basis, who approved it, and what interval followed.
The chain nearly always breaks in one place. A procedure revised three months into the campaign, with no record of which revision governed which reading. A technician whose certification expired on the fourteenth while the outage ran to the twenty-second. A post-job calibration check that was performed but never written down. Any one of these turns a sample finding into a systemic one, because if the link cannot be shown for the sampled reading, the auditor has no basis to assume it exists for the other four thousand.
This is why audit readiness is a data model question rather than a filing question. Sites that pass are not measuring more carefully. They are storing the circumstances of the examination alongside the result, at the moment the result is captured, so reconstructing any reading takes seconds rather than a week of pulling paper. Once that exists, the audit stops being an event to prepare for and becomes a query. The preparation effort that used to consume six weeks before every visit stops being needed, which is a larger saving than the audit itself.
Flow-accelerated corrosion is what the grid is actually watching
In a gas or combined-cycle plant, the mechanism that most often drives a CML programme is flow-accelerated corrosion. Carbon steel in high-purity water forms a protective magnetite film; under the right combination of flow velocity, geometry, temperature and reducing water chemistry, that film dissolves faster than it reforms, and the metal underneath goes with it. It is not a pitting mechanism and not general corrosion. It produces smooth, scalloped, sharply bounded thinning immediately downstream of geometry that disturbs flow: elbows, tees, reducers, orifice plates, control valve outlets, feedwater heater drain lines.
Single-phase attack lives in feedwater, economizer inlet piping and the associated headers, peaking in a temperature band well below full steam conditions. Two-phase attack lives in LP evaporator circuits, moisture separator and heater drains, deaerator inlets and cascading drain systems, where the water and steam mixture changes the local chemistry. Susceptibility is dominated by alloy content: small chromium residuals in the steel dramatically slow the mechanism. That means two elbows of the same specification, installed on the same day, from two different heats, can thin at rates that differ by a factor of several. It also means heat traceability and material verification are integrity inputs rather than construction paperwork.
The registry consequences are specific. Store the chromium content or the material verification result at the component where you have it, because it is the strongest single predictor you will ever hold. Store the geometry class of the location, since downstream-of-elbow behaves nothing like a straight run. Store the water chemistry regime in force during the interval, because a change in treatment can move rates across an entire circuit at once and you will want to see the step change rather than argue about it. Then a rising rate becomes an explainable event rather than a surprise found during a walkdown.
A condition monitoring location is often not a thickness location
The abbreviation CML is broader than TML, and in power generation the difference is not academic. Main steam and hot reheat piping in a modern combined-cycle unit is creep strength enhanced ferritic material, typically Grade 91. Its failure modes are creep damage, thermal fatigue and, in badly heat-treated welds, Type IV cracking in the fine-grained heat affected zone. None of that shows up as wall loss. It is monitored by hardness survey, metallographic replication, weld examination and dimensional checks for creep swelling, and the results are hardness values, damage classifications and crack indications rather than millimetres.
Superheater and reheater tubing adds another non-thickness measurement: steam-side oxide scale thickness, measured ultrasonically, used to infer the tube metal temperature history and estimate creep life consumption. Attemperator downstream piping needs cracking examination rather than wall measurement. Drum nozzles and header ligaments need fatigue-focused examination tied to the start count. All of these are monitoring locations that belong in the same programme, with the same interval logic and the same evidence chain, and all of them are excluded by a data model that assumes every location produces a thickness.
What happens in practice when the registry cannot hold them is that they migrate into separate files kept by whoever owns that speciality. The hardness surveys sit in a consultant's report. The oxide scale results sit in a tube programme spreadsheet. The registry looks complete while covering perhaps sixty percent of the actual monitoring programme, and the audit finds the gap by asking a single question about a Grade 91 weld. A registry that accepts a typed result at a monitoring location, with its own units, acceptance basis and trend, keeps the programme in one place and makes the coverage claim honest.
Cycling changes the denominator: starts, not years
Combined-cycle plants that were designed as baseload machines now two-shift. A unit that once ran continuously through a season may now start daily, follow load through the afternoon and shut down overnight. The damage accumulation profile changes completely, and the annual rate that has been reported for a decade quietly stops describing anything real. Wall thinning by flow-accelerated corrosion continues to track operating hours and water chemistry, so an hours denominator remains defensible. Thermal fatigue does not care about hours at all.
Fatigue damage at an attemperator downstream spool, a drum nozzle or a header ligament accumulates per thermal transient, weighted by ramp rate and the temperature swing achieved. A unit that went from four hundred starts a year to eleven hundred did not triple its corrosion rate, but it may well have tripled the rate of a fatigue mechanism at a handful of specific locations. Reporting both against a calendar interval mixes the two and produces a picture that is directionally wrong for both. Worse, it hides the correlation you most want to see: the location where an inspection finding followed a period of unusually aggressive starts.
The registry fix is unglamorous and effective. Store operating hours and start counts for the interval between examinations, pulled from the historian rather than typed, and attach them to the interval rather than to the year. Let each mechanism nominate its denominator. Then a wall loss result reads in millimetres per thousand operating hours where that is what matters, a fatigue location reads against accumulated starts, and both can still be expressed per year for the report the boiler inspector expects. Nothing is lost, and the analysis finally matches the way the unit is being run.
Mill tolerance and the phantom first-interval corrosion rate
A baseline error is the most common arithmetic trap in power thickness programmes, and it is entirely avoidable. Pipe ordered to nominal wall thickness under SA-106 is permitted a negative tolerance of twelve and a half percent and is fully compliant at that thickness. Tubing ordered to minimum wall under SA-213 has no negative tolerance, but tubing ordered to average wall does. So a schedule 80 line whose nominal wall is 7.62 millimetres may leave the mill at 6.67 millimetres and be exactly what was purchased.
Baseline the corrosion calculation against nominal, and the first inspection appears to show almost a millimetre of loss that never occurred. Divide by the interval and the rate is fictitious, remaining life collapses, and a plant either replaces good pipe or, more commonly, learns to distrust its own trend data and starts ignoring the alarms. The opposite error also happens: a component that genuinely lost wall shows a modest rate because someone quietly adjusted the baseline downward to make the first result look reasonable, with no record of the adjustment.
The correct baseline is a measured as-installed thickness at the exact monitoring location, taken during construction or at the first opportunity, stored as its own field and never confused with nominal. Where no as-installed reading exists, the honest approach is to declare the first inspection as the baseline, accept that no rate exists until the second inspection, and record that state explicitly rather than manufacturing a rate from nominal. Auditors respond well to a system that says it does not yet know, and badly to one that produced a confident number from an assumption nobody documented.
Who inspects a merchant combined-cycle plant, and what each one wants
The jurisdictional inspector, commissioned through the National Board or the state boiler authority, is concerned with the boiler and its appurtenances under ASME Section I, and with any repair or alteration performed under the National Board Inspection Code. They want to see that repairs were done under an accepted programme, that the inspector was involved where required, and that the records exist. Thickness data enters their view when it supports a fitness decision or justifies deferring work.
The insurer's inspector, and the boiler and machinery underwriter behind them, want evidence that the inspection programme is being executed rather than merely written. They tend to ask about coverage: which locations were scheduled, which were actually examined, and what happened to the ones that were missed because scaffolding was not available. The lender's independent engineer, on a project-financed plant, asks a version of the same question with a financial frame, since deferred inspection becomes a deferred capital risk in their model. An original equipment manufacturer running a long-term service agreement asks whether the inspections their agreement requires were done to their scope.
Each of these produces a different evidence pack from the same underlying records, and the recurring failure is that all of them are assembled by hand. A registry that can filter by system, date range, mechanism and completion status, and export a coherent pack with the supporting certificates attached, turns four separate multi-week exercises into four queries. It also exposes the metric these audiences care about most and that spreadsheets rarely show honestly: schedule adherence, meaning the fraction of scheduled locations actually examined in the window, with reasons recorded for those that were not.
Getting audit-ready when the audit is six weeks away
With a fixed date approaching, sequence matters more than scope. Start with the sample the auditor will take. Pick twenty readings at random from the last two years, across at least three systems, and try to reconstruct each one completely: procedure revision, technician certification on the date, instrument calibration and checks, calculation inputs, approval. Whatever fraction you can reconstruct in an hour is a fair estimate of what the audit will find. That exercise usually identifies the one or two broken links that matter, and it takes a morning.
Then fix the links rather than the volume. If certification validity is the weak point, load the certificates with their expiry dates and turn on validation at capture, so the problem cannot recur even if the historical gap remains. If procedure revision is the weak point, stamp the revision onto every reading going forward and reconstruct the revision history for the audit window from the document control system. Historic gaps should be documented as known gaps with a stated corrective action, which is a far stronger position than a gap discovered by the auditor.
Finally, rehearse the export. The pack you hand over should assemble itself from the registry, contain the certificates and calibration records as attachments rather than as references to another system, and read in the order the auditor works. If the first pack has to be produced by an engineer working evenings, the system has not actually solved the problem, and next year the same evenings will be needed again. Atlantis NDT can run that reconstruction exercise against a sample of your own records as part of a working demonstration. Write to info@atlantisndt.com to arrange a consultation or request a quote.
Why do thickness audits fail on traceability rather than on the numbers?
Because the number is rarely in dispute. The auditor samples a reading and follows it backwards: procedure revision, technician qualification on that date, instrument calibration, the calculation and the approval. The chain usually breaks in one link, and that single break puts the whole dataset in question rather than the one reading. Sites that pass are not measuring more accurately than sites that fail. They are simply able to reconstruct the circumstances of any reading in front of the auditor, within minutes.
What is flow-accelerated corrosion actually doing to a combined-cycle plant?
It dissolves the protective magnetite layer on carbon steel where flow, temperature and water chemistry combine badly, then removes the metal underneath. Single-phase attack concentrates in feedwater and economizer inlet geometry, two-phase attack in LP evaporator circuits and heater and separator drains. Chromium is the strongest defence: even small residual amounts sharply reduce susceptibility, which is why two identical elbows from different heats can thin at completely different rates. That makes heat traceability an integrity variable, not a purchasing detail.
Is a condition monitoring location always a thickness location?
No, and treating the two as synonyms is a common design mistake. On Grade 91 main steam and hot reheat piping the damage of interest is creep and thermal fatigue, monitored through hardness surveys, replication and weld examination rather than wall loss. On superheater and reheater tubing, steam-side oxide scale measurement informs tube metal temperature history. A registry has to accept a non-thickness result at a monitoring location and trend it in its own units, or those locations end up in a separate file nobody audits.
How does daily cycling change the way corrosion rates should be calculated?
It splits the damage into two clocks. Wall thinning tracks operating hours and water chemistry, so an hours-based denominator is defensible. Thermal fatigue at attemperators, drum nozzles and header ligaments tracks starts and ramp rates, so a per-year figure is meaningless for a unit that went from four hundred starts a year to eleven hundred. Store operating hours and start counts against each interval, pull both from the historian, and let each mechanism use the denominator that fits it.
What is the phantom corrosion rate that appears at the first inspection?
It comes from using nominal wall as the baseline. Pipe ordered to nominal wall under SA-106 can be up to twelve and a half percent under nominal and still be fully compliant, while tubing ordered to minimum wall under SA-213 has no negative tolerance. Baseline from nominal and the first inspection appears to show years of loss that never happened. The correct baseline is a measured as-installed thickness at the same location, recorded once and carried forward as a distinct field.
Is API 510, 570 or 653 inspector training part of this offer?
No. Atlantis NDT provides NDT method training to ASNT SNT-TC-1A and ISO 9712, at Level I, II and III across UT, RT, MT, PT, ET, VT, PAUT and TOFD, alongside ASNT Level III consulting, inspection management and reporting software, digital twins, 3D laser scanning and report validation. API inspector certification programmes are run by API and fall outside that scope. The registry records and evidences work performed under those codes; it does not certify anyone.
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.