Standardising remaining life and inspection intervals across a combined-cycle fleet
In a combined-cycle fleet the interval engine's hardest job is not the arithmetic, it is provenance. Sites derive dates from different clocks — calendar years, fired hours, equivalent starts, jurisdictional certificate dates, insurer schedules and OEM factored-hours tables. Until every due date carries the rule that produced it and the clock it ran on, a fleet backlog number cannot be defended.
A refinery integrity engineer moving into power generation discovers quickly that API 510 and 570 are not jurisdictional here. Main steam, hot reheat and high-pressure feedwater are ASME B31.1 power piping, and B31.1's Chapter VII obliges the operating company to run a documented programme for covered piping systems. The boiler itself is inspected under the National Board Inspection Code and the state's rules, by a commissioned inspector, on a certificate cycle the jurisdiction sets. Meanwhile the fuel gas yard may be B31.3, the gas turbine runs to the OEM's factored fired hours and factored starts, and the property insurer has its own schedule that in practice moves the outage. Five clocks, one plant. Multiply by twelve plants in seven states and the reason the fleet numbers do not roll up is not sloppiness — it is that nobody ever wrote down which clock each site was using.
Source: Rules referenced here: ASME B31.1 Power Piping, including the Chapter VII operation and maintenance provisions covering covered piping systems; ASME B31.3 Process Piping for balance-of-plant process lines; ASME Boiler and Pressure Vessel Code Sections I, V and IX; the National Board Inspection Code NB-23 Part 2 for in-service inspection and repair; state jurisdictional boiler and pressure vessel rules and the National Board commissioned inspector regime; EPRI guidance on flow-accelerated corrosion and HRSG cycle chemistry; and API 510, API 570 and API 579-1/ASME FFS-1 where an owner has voluntarily adopted them for balance-of-plant equipment.
| System | Dominant degradation | Clock the interval should run on | Rule or authority that sets it |
|---|---|---|---|
| HP superheater and reheater headers, Grade 91 piping | Creep and Type IV cracking in weld heat-affected zones | Time at temperature plus starts; not a wall thickness trend | B31.1 Chapter VII covered piping systems programme |
| Feedwater, condensate and LP evaporator circuits | Flow-accelerated corrosion | Operating hours inside the susceptible temperature and chemistry band | Owner FAC programme, typically supported by predictive modelling |
| Attemperator downstream piping and spray stations | Thermal fatigue from spray events and rapid load change | Number of spray events and unit starts | Owner programme informed by OEM and industry guidance |
| Drums and pressure parts under the boiler certificate | General corrosion, pitting, under-deposit corrosion | Jurisdictional certificate cycle | NBIC NB-23 Part 2 and the state boiler authority |
| Gas turbine and hot gas path | Oxidation, thermal-mechanical fatigue, coating loss | Factored fired hours and factored starts | OEM maintenance interval tables |
| Fuel gas, lube oil and balance-of-plant process piping | General and localised corrosion | Calendar time with a corrosion-rate derivation | B31.3 practice; API 570 methodology only where the owner adopts it |
Why the same component gets three different due dates in one fleet
Ask a fleet integrity manager for the total number of overdue inspections across twelve combined-cycle plants and you will usually get a number. Ask how it was assembled and the number falls apart. One site counts an item overdue when the jurisdictional certificate date passes. Another counts it overdue against the OEM manual. A third has a spreadsheet of API 570 circuits set up years ago by an engineer who came from a refinery, and counts against half-life intervals nobody has revisited. A fourth counts nothing overdue because its insurer's last visit closed everything out.
None of these sites is being careless. Each is applying a rule that is legitimate for its own context. The problem is that the rule is not written down anywhere the fleet can see, so the four numbers cannot be added. When a corporate integrity review, a regulator or an incoming owner asks for a consolidated backlog, the answer takes six weeks of phone calls and still comes with a caveat, because the underlying dates were produced by four different logics against four different clocks.
This is the standardisation problem in power generation, and it is a data model problem before it is a policy problem. You cannot legislate a single interval rule across a fleet of units that differ in jurisdiction, insurer, configuration and duty. What you can do is require that every due date in the fleet carries the rule that produced it, the clock that rule ran on, the parameters that fed it, and the person accountable for the rule. Once that field exists, comparison becomes possible and disagreements become specific.
The clock is equivalent operating hours, not the calendar
A combined-cycle plant in 2026 does not run the way the equipment was specified to run. Units built for baseload duty now cycle daily against renewable output, with two starts a day in some markets and long low-load holds in others. That change is not cosmetic. Thermal fatigue damage at attemperator spray stations, header ligaments, drum nozzle penetrations and superheater tube-to-header welds accumulates per thermal transient, not per year of ownership.
Gas turbine OEMs solved the notation problem long ago by expressing maintenance intervals in factored fired hours and factored starts, where a start consumes a defined number of equivalent hours. The fixed-equipment side of the plant rarely follows suit, and that is where fleets lose comparability. Site A tracks fired hours because the turbine team's system does. Site B tracks starts. Site C tracks calendar because that is what the certificate cycle uses. Three plants, three units of measure, one impossible roll-up.
The remedy is to let the interval engine hold the clock as an explicit attribute of the rule. An interval of forty-eight months is a different commitment from an interval of twenty-four thousand equivalent operating hours, and both are legitimate as long as the system knows which one it is applying and can convert between them using a documented start factor. During an evaluation, ask the vendor to configure one rule on a calendar clock and one on an equivalent-hours clock, then change the unit's operating profile and watch which dates move.
A thickness trend will not see the failure that closes the plant
The failures that take combined-cycle units out for months are not usually wall loss failures. They are creep and creep-fatigue in Grade 91 and Grade 22 headers and hot reheat piping, Type IV cracking in the fine-grained heat-affected zone of girth welds, thermal fatigue cracking downstream of attemperators, and tube failures in the HRSG. In almost all of those, the wall thickness at the failure location is close to as-built at the moment of failure.
This matters commercially because integrity software is usually sold on its thickness module, and thickness is what fleets are able to load first. A system that computes remaining life from a thickness trend will happily publish a reassuring number for a Grade 91 header that is halfway through its creep life. Nothing in the arithmetic is wrong; the arithmetic is simply answering a question that has no bearing on how the component will fail.
So a fleet standard has to classify components by governing damage mechanism before it standardises intervals, and the software has to support more than one kind of interval basis on one asset register. Creep-limited components need time at temperature, transient counts and targeted weld examination on a defined schedule. Metal-loss components need a rate and a required thickness. Mixing them under one remaining-life formula produces a tidy dashboard and a false picture, which is the worst combination available.
Flow-accelerated corrosion defeats a fixed monitoring grid
Flow-accelerated corrosion is the metal loss mechanism that actually matters in the water and steam cycle, and it does not behave like the general corrosion the API formulas were built around. It is single-phase or two-phase, it peaks in a particular temperature window, it is strongly sensitive to pH, dissolved oxygen and the chromium content of the steel, and it concentrates immediately downstream of geometric disturbances: elbow extrados, tee runs, orifice plates, reducers and control valve outlets.
A monitoring grid laid out on straight pipe at fixed spacing will miss it entirely, and the resulting remaining life will be long and wrong. What the software has to do is hold the susceptibility screening and the predictive model output alongside the measured data, so that inspection scope is selected from the model rather than from where somebody could reach with a ladder five years ago. When the model changes — new chemistry regime, replaced component in a chromium-bearing grade, a changed operating profile — the scope should change with it.
There is a fleet dimension here too. FAC programmes vary enormously in maturity across a fleet, and the difference is usually invisible in a backlog report because both a mature and an immature programme report a number. A useful evaluation question is whether the system can show, per unit, the proportion of FAC-susceptible components that have ever been examined and when. That single figure separates the sites doing the work from the sites reporting on work they have not scoped.
One rule library, per-site parameters, and a rule owner
The architectural decision that determines whether standardisation succeeds is where rules live. In fleets that fail, each site holds its own rules because the software's configuration is per-site, and the fleet standard exists only as a document that sites are meant to follow. Within two years the configurations have drifted, nobody can say by how much, and the standard document is a historical artefact.
In fleets that succeed, the rule library is fleet-level and singular. It defines what a covered piping system is, what qualifies as a Class 1 circuit for an owner who adopted API methodology, what the start factor is in the equivalent-hours clock, what the FAC susceptibility screen looks like, and what caps apply. Each rule carries a version, an effective date and a named owning engineer. Sites hold parameters — jurisdiction, insurer, unit configuration, outage windows, operating profile — and parameters are exactly the things that should differ between sites.
The practical test is a permissions test rather than a feature test. Ask the vendor to show you what a site engineer can change and what they cannot. If a site user can edit the interval formula rather than the input to it, the fleet number has an expiry date measured in months. If a site user can only propose a change that routes to the rule owner for approval, with the change recorded against the rule version, then standardisation is a maintainable state rather than a one-off cleanup project.
Standardising the data model before standardising the intervals
Fleets almost always try to standardise intervals first, because intervals are what the executive asked about. It rarely works, because two sites cannot agree on an interval for a component they have not agreed on the identity of. One plant's asset register has the HP evaporator as a single functional location. Another has it broken into eleven components. A third holds the HRSG as one tag and keeps the internals in a drawing set nobody has digitised.
So the first standardisation deliverable is a component taxonomy that every unit maps into: what is an asset, what is a circuit, what is a component, where a covered piping system boundary sits, and how a monitoring location is identified so it can be re-found by a different technician on a different contract in five years. The taxonomy is unglamorous and it is the whole job. Once it exists, interval rules attach to component classes and standardisation becomes a configuration exercise rather than a negotiation.
Expect the mapping to expose real inconsistency, and plan for it. It is common to find that two nominally identical units have different piping arrangements because they were built four years apart by different EPCs, or that a component the fleet standard assumes exists was never installed at one site. Discovering that during a mapping exercise is cheap. Discovering it when a fleet-wide inspection scope is issued and one site cannot execute it is not.
Making a fleet number that survives its first challenge
The output of all this is a small set of numbers that a fleet integrity manager can put in front of an executive, a regulator, an insurer or a buyer's technical diligence team without a caveat. Backlog by unit and by criticality. Schedule compliance against the derived dates. The proportion of covered piping system scope examined in the current programme cycle. The proportion of FAC-susceptible components with a current examination. Deferrals open, aged, and with an engineering basis attached.
Each of those numbers will be challenged the first time it is published, and the challenge is always the same in form: that number is wrong for my unit, because you counted something I do not count. The only durable answer is to drill from the fleet number to the unit, to the component, to the derived date, and to the rule, version and owner behind it. If the drill-down exists, the challenge resolves in ten minutes and usually improves the rule. If it does not, the number is quietly abandoned and the fleet returns to phone calls.
This is why provenance, not arithmetic, is the buying criterion in a multi-site power fleet. Every candidate system can divide a thickness by a rate. Very few can tell you, two years later, which rule version produced a specific date, what the parameters were at the time, who approved the deviation, and what the number would have been under the fleet standard rather than the site's local practice.
What to test during an evaluation
Take two units that genuinely differ — one baseload, one cycling hard, ideally in different jurisdictions — and load both into the candidate system with their real asset registers, not a cleaned sample. Then ask for a single consolidated overdue list. The interesting part is not whether a list appears; it is whether the system forces you to declare, per component class, which rule applies, and whether it refuses to publish a fleet number until every component has a rule attached.
Next, run the drift test. Change one site's operating profile from baseload to two starts a day and see which dates move. If nothing moves, the system is on a calendar clock everywhere and the fleet number will misstate the cycling unit permanently. Then change a rule at fleet level and check that every affected date across both units recalculates, that the previous dates remain visible in history, and that anyone who had been working to the old date is notified rather than left with a stale printout.
Finally, test the awkward integrations. The jurisdictional certificate date lives with the plant's regulatory coordinator. The turbine's factored hours live in the OEM's system or in the control historian. The insurer's schedule arrives as a letter. Ask how each of those reaches the interval engine, and who is accountable when one of them is stale. A system that can only be right when all three feeds are manually maintained by one person at head office has not solved the standardisation problem, it has relocated it.
Why do two identical HRSGs in the same fleet show different intervals?
Usually because they were commissioned by different people under different assumptions. One site inherited an integrity engineer from refining and set up API 570 circuits with half-life intervals. Another follows the OEM manual. A third does only what the jurisdictional inspector requires for the certificate, and a fourth follows its insurer's recommendations. All four are defensible in isolation. None of them roll up, because the numbers were produced by different rules against different clocks and no field in any system records which.
What is an equivalent operating hours clock and why does it matter?
It is a duty counter that weights starts against running hours, so that a unit cycling twice a day accumulates damage faster than a baseload unit with the same fired hours. Gas turbine OEMs have used factored fired hours and factored starts for decades for exactly this reason. Thermal fatigue at attemperators, headers and drum penetrations behaves the same way. An interval derived on calendar years understates damage at a peaker and overstates it at a baseload unit in the same fleet.
Can a thickness-based remaining life cover a Grade 91 header?
No, and treating it as though it can is the most dangerous single error in this space. Creep damage in Grade 91 accumulates as cavitation and eventually Type IV cracking in the fine-grained heat-affected zone of a weld, at a wall thickness that is essentially unchanged. A thickness trend on that header will show a comfortable remaining life right up to the point of failure. Creep-limited components need a time-at-temperature assessment and targeted weld examination, tracked on a different clock entirely.
How should flow-accelerated corrosion components be handled differently?
FAC wall loss is local, geometry-driven and concentrated downstream of elbows, orifices, tees and reducers, inside a specific temperature and chemistry band. A fixed four-point CML grid on a straight run will miss it. The system needs to hold the FAC susceptibility screening, the predictive model output and the measured grid together, so the inspection scope follows the model rather than a legacy grid, and so a change in cycle chemistry or feedwater pH visibly changes the wear rate and the next examination.
Who owns an interval rule once a fleet standardises?
One named engineer per rule, with a version history, not a committee and not a site. The rule library is fleet property: the definition of a Class 1 circuit, the FAC susceptibility screen, the CPS scope boundary, the start factor used in the equivalent hours clock. Sites hold parameters, not rules — jurisdiction, insurer, unit configuration, outage dates. The moment a site can edit the rule rather than the parameter, the fleet number stops being comparable and nobody notices for two years.
Is API 510, 570 or 653 inspector training part of this offer?
No. Atlantis NDT delivers NDT method training to ASNT SNT-TC-1A and ISO 9712, Level I, II and III across UT, RT, MT, PT, ET, VT, PAUT and TOFD, together with ASNT Level III consulting, inspection management software, reporting software, digital twins, 3D laser scanning and report validation. API inspector certification is administered by the American Petroleum Institute through its own individual certification programme. The software applies interval rules; it does not certify the people who approve them.
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.