One corrosion rate method for every plant in the group
Corrosion rate calculation in chemical manufacturing computes a short-term rate from the last two thickness readings and a long-term rate from the original baseline, then lets the more conservative of the two govern remaining life. Standardising it across sites means fixing one t-min basis, one baseline rule and one noise threshold, so a 12 mpy rate means the same thing everywhere.
API 510 §7.1.1 and API 570 §7.1.1 both define the two rates the same way: long-term uses the initial thickness and the current reading over the full elapsed interval; short-term uses the previous reading and the current one. Neither code tells you which is right, and that is the whole problem in a multi-site group. One plant takes the higher rate automatically. Another lets the inspector pick, and the picks are not documented. A third clamps negative short-term rates to zero, which quietly deletes the evidence that the readings disagree. Chemical service makes the divergence worse, because campaign plants expose a vessel to HCl or concentrated sulphuric for a few hundred hours a year and then run something benign, so a calendar-year rate understates the aggressive service by an order of magnitude. Roll-up only works when the exposure basis, the t-min basis and the outlier rule are identical.
Source: Basis: API 510 Pressure Vessel Inspection Code §7.1.1 (corrosion rate and remaining life) and §8.2 (inspection intervals); API 570 §7.1.1 and its piping class interval table; API 571 damage mechanism descriptions; API 584 Integrity Operating Windows; ASME Section VIII Division 1 UG-27 for pressure-design thickness; ASTM A106 and ASTM A20 wall and plate tolerances; OSHA 29 CFR 1910.119(d)(3)(ii) and (j) for RAGAGEP and mechanical integrity.
| Decision point | How one site does it | How the other site does it | Effect on the computed rate | Basis to standardise on |
|---|---|---|---|---|
| Baseline thickness | Nominal thickness taken from the drawing | First measured reading at commissioning | Nominal against material under-tolerance can invent or hide 0.010–0.030 in of apparent loss | First measured reading, with nominal retained only as a labelled fallback |
| Minimum required thickness | Design thickness minus the stated corrosion allowance | UG-27 pressure-design thickness at the current rated pressure | Remaining life differs by years on identical measured metal | Code-calculated t-min at the current rating, stored with the calculation that produced it |
| Negative short-term rate | Clamped silently to zero | Left negative and averaged into the trend | Clamping hides a measurement dispute; averaging drags the governing rate down | Flag as a data-quality exception, exclude from the governing rate, trigger a re-read |
| Governing rate selection | Always the higher of the two rates | Inspector judgement, not recorded anywhere | Two sites publish different remaining life from the same readings | Higher rate governs by default; any override carries a written basis and a named approver |
| Exposure clock | Calendar years between readings | Calendar years between readings | Campaign equipment understates its rate whenever aggressive service is intermittent | Calendar rate plus a service-hours rate for campaign equipment, both stored and both visible |
| Lined and clad equipment | Thinning rate computed on total measured wall | Location excluded, with no reason recorded | Overlay thickness masks base-metal loss, or the asset drops out of the programme entirely | Explicit not-rate-governed classification that names the alternative inspection method |
The job: five sites, five spreadsheets, no common column
The trigger is almost never a corrosion problem. It is a meeting. Corporate integrity asks for the twenty shortest remaining lives across the group, five plants send five spreadsheets, and the numbers cannot be put in one table. One site reports years to minimum thickness. Another reports years to the next turnaround. A third has been publishing the long-term rate only, because the short-term rate went erratic when they changed ultrasonic contractors. Nothing here is dishonest. Every site follows a procedure that was written locally and has served that site well for fifteen years.
What makes it hard to fix is that the disagreement is not in the arithmetic. Everyone computes the previous thickness minus the current thickness, divided by the interval. The disagreement lives in the inputs and in the rules wrapped around the arithmetic: which thickness counts as the baseline, what minimum thickness remaining life is measured against, what happens when the short-term rate comes out negative, whether the governing rate is selected automatically or by an inspector, and whether that selection is recorded anywhere. Six small decisions, each individually defensible, produce remaining lives that differ by years on the same vessel.
Standardising is therefore a data-modelling problem before it is a software problem. The system has to hold each site's practice explicitly rather than burying it inside a formula, so a group-level view can state which basis produced a number, and so moving a site onto the corporate basis is a dated, visible, reversible event rather than a silent recalculation of twenty years of history.
What the inspection codes fix, and what they deliberately leave open
API 510 and API 570 both define the two rates identically and both express remaining life as the current thickness minus the required thickness, divided by the governing corrosion rate. Both then bound the next inspection at half the remaining life, subject to a code maximum — ten years for internal or on-stream vessel inspection under API 510, and the class-dependent thickness measurement intervals in API 570 for piping. That structure is not negotiable, and it means the corrosion rate is not a reporting statistic. It is the input that sets when someone opens the equipment.
What the codes leave open is the choice between the two rates. The inspector is expected to review both and apply the one that best represents current conditions, which is sound engineering and terrible governance if the reasoning is never captured. A group that has not written down a default and an override mechanism has, in practice, as many methods as it has inspectors. The codes also permit an estimated rate where no history exists, drawn from similar service, owner-user experience or published data, confirmed by an on-stream measurement after roughly a thousand hours of service.
Under OSHA process safety management, this stops being a housekeeping matter. The mechanical integrity element requires inspections and tests to follow recognised and generally accepted good engineering practice, and requires the employer to document that equipment complies with it. If five sites hold five interpretations of the same code, an auditor will reasonably ask which one is the company's, and the answer needs to be a document rather than a habit.
Measurement noise sets the floor on any short-term rate
Point-to-point ultrasonic repeatability on corroded carbon steel is realistically around five thousandths of an inch once probe, couplant, surface condition and technician are all allowed to vary. Over a twelve-month interval that scatter alone produces roughly five mils per year of apparent corrosion. Over six months it produces ten. On equipment that is genuinely losing three to eight mils per year, a short-term rate taken on a one-year interval is at least half noise, and the sign of the result is close to a coin toss.
Temperature makes it worse in a way that is easy to miss because it is systematic rather than random. Ultrasonic velocity in carbon steel falls by roughly one percent per hundred degrees Fahrenheit of temperature rise, so an uncompensated on-stream reading at 500 °F reads about two and a half percent thin. On a half-inch wall that is twelve thousandths of an inch of fictitious loss, appearing as a twelve mil-per-year spike in the year the reading was taken on-stream and an equally fictitious negative rate the year the unit was down and cold.
The practical response is a detection limit rather than a suppression rule. The system should refuse to publish a short-term rate when the measured difference is smaller than the configured resolution for that interval, report it as below detection, and carry the long-term rate for remaining life. Choosing whether the governing point is the grid minimum or the grid mean matters here too: the minimum is an extreme value and is systematically more volatile than the mean, so two sites using different statistics will not agree even before they disagree about anything else.
Chemical service breaks the calendar-year assumption
A refinery unit runs the same fluid for four years. A chemical plant runs eleven products through the same reactor train. Corrosion in that train is a function of which campaign was running, not of how many months have elapsed, and a calendar-based long-term rate averages the aggressive service into the benign service until the number stops describing anything. A vessel in hydrochloric service four hundred hours a year can carry a calendar rate of three mils per year and an in-service rate approaching sixty. Both are arithmetically correct. Only one of them tells you what happens if the campaign schedule doubles.
Several common chemical services are also governed by a variable that has nothing to do with time. Carbon steel survives concentrated sulphuric acid only because of a protective ferrous sulphate film that is stripped above a modest velocity threshold, so the same metal in the same acid corrodes at a fraction of a mil per year or at hundreds depending on flow. Caustic service is bounded by a concentration-and-temperature envelope beyond which the mechanism switches from thinning to caustic stress corrosion cracking. In neither case does a thickness trend give warning.
This is why a chemical corrosion rate module needs to carry an exposure basis and a link to integrity operating windows in the sense of API 584. The useful behaviour is not a prettier trend chart. It is that an excursion outside a defined window — a velocity, a concentration, a temperature, a chloride level — raises an inspection action immediately, rather than waiting for a scheduled reading that will average the excursion away.
Baselines move, and long-term rates move with them
The long-term rate is only as good as the thickness it started from, and that starting point moves more often than most registers record. A spool is replaced during a turnaround and the location keeps its old identifier. The new pipe arrives at 0.322 in against a baseline of 0.280 in measured on the old pipe, the long-term rate turns negative, and the site's clamp-to-zero rule quietly converts a data event into a zero corrosion rate on a component that has never actually been assessed.
Using nominal thickness as the baseline introduces a different error, and it is one-sided. Seamless pipe to ASTM A106 is permitted a wall under-tolerance of twelve and a half percent, so a nominal 0.280 in schedule 40 line can legitimately leave the mill at 0.245 in. A long-term rate computed from nominal invents thirty-five thousandths of loss that never happened, and on a twenty-year-old line that is enough to dominate the answer. Clad and weld-overlaid vessels invert the problem: total measured wall includes an overlay that carries no pressure, so a rate computed on total thickness understates the loss of the base metal that actually matters.
The corporate rule worth adopting is simple to state and requires real software to enforce. The baseline is the first measured reading. Any change to a baseline — a replacement, a repair, a re-clad, a change of measurement point — is a dated record with a reason and an author, and the previous history is preserved rather than overwritten. That single discipline removes most of the cases where two sites report different rates from what appear to be identical datasets.
Equipment where a thinning rate is the wrong instrument entirely
Chemical manufacturing runs materials that a thickness-based programme cannot describe. Glass-lined reactors, rubber and PTFE-lined vessels, brick-lined columns and FRP ductwork do not fail by general thinning; they fail through a holiday, a permeation path or a liner disbondment, and the correct examination is spark testing, holiday detection or thermography rather than an ultrasonic series. Zirconium, tantalum and high-nickel alloys in aggressive acid duty are usually specified precisely because their general corrosion rate is negligible, so the number the system computes will be noise and the real risk is localised.
Whole damage mechanism families are invisible to a rate as well. Chloride stress corrosion cracking of austenitic stainless above roughly 140 °F, caustic cracking, amine cracking, sigma phase embrittlement in 300-series steels held at high temperature, and hydrogen attack in hydrogen service all progress with no measurable wall loss. A corrosion rate module that only understands thinning will report a 316L line as comfortably healthy right up to the moment it leaks through a crack, and will do so with a green indicator.
The practical requirement is an explicit not-rate-governed classification that names the alternative method and keeps the equipment inside the plan. This matters more than it sounds. In most chemical plants the single most common way equipment falls out of an integrity programme is a blank rate field: with no rate there is no remaining life, with no remaining life the item never appears on the shortest-life report, and an asset that is never on a report is an asset nobody is looking at.
How to evaluate a corrosion rate module across a multi-site group
Ask for demonstrations rather than feature lists. Show me the same vessel computed simultaneously on two different minimum-thickness bases. Show me a negative short-term rate and tell me what the system does with it without a human intervening. Show me the audit record of a governing-rate override, including the approver and the technical basis. Show me what happens to twenty years of history when a site changes its baseline rule — specifically, whether the previously published numbers survive as a record or are silently recomputed.
Then test the group-level questions, because that is the job you are actually buying for. Can the rate be expressed per exposure hour as well as per calendar year, and can both appear on the same report? Can a measurement location be marked not-rate-governed and still appear on the inspection plan with its own method and interval? Does the interval calculation apply both half remaining life and the code maximum, and does it state which of the two bound the answer? Can a corporate standard be published while a site retains a documented, dated exception?
Atlantis builds this as a configured Odoo 18 application rather than a fixed product, because the whole point is that your corporate basis is yours and the per-site exceptions have to be modelled honestly during transition. It is affordable, accessible and fully customizable, and the specification work is done with an ASNT Level III who has run these programmes. If you want to see it against your own data, ask for a working session at info@atlantisndt.com and bring one vessel that two of your sites disagree about.
How is the short-term corrosion rate different from the long-term rate?
The long-term rate uses the baseline thickness and the current reading across the whole elapsed period, so it averages every process change the equipment has ever seen. The short-term rate uses only the previous reading and the current one, so it reflects recent service but inherits the full measurement error of two readings across a shorter interval. Long-term is stable and slow to react; short-term is responsive and noisy.
Why does the more conservative rate govern instead of an average?
Averaging the two assumes both describe the same process, and they do not. If the short-term rate is higher, something changed recently and an average would delay recognising it. If the long-term rate is higher, past service was harsher and an average credits the equipment with an improvement it may not keep. Taking the higher value is a deliberate bias toward earlier inspection, which is exactly what a remaining-life calculation is for.
How many years of readings do you need before a short-term rate is real?
Divide your ultrasonic repeatability by the interval. At roughly five thousandths of an inch of point-to-point scatter, a one-year interval carries about five mils per year of apparent corrosion that is purely measurement. On equipment genuinely corroding at three to eight mils per year, that means a two to four year interval before the short-term rate carries more signal than noise. Shorter intervals need a stated detection limit.
Does a campaign plant need a different corrosion rate basis?
Yes. A vessel exposed to hydrochloric or concentrated sulphuric service for four hundred hours a year and to benign service the rest of the time has a calendar rate roughly twenty times lower than its in-service rate. If the campaign schedule changes, the calendar rate becomes wrong immediately and stays wrong until the next reading. Storing exposure hours alongside calendar time lets both numbers come from the same readings.
What breaks when two sites use different minimum-thickness definitions?
Remaining life breaks, and with it the inspection interval, since the interval is bounded by half the remaining life. Design thickness minus corrosion allowance is usually thicker than the pressure-design thickness calculated for the current rated pressure, so one site re-rates or retires equipment years before another would on identical metal. Group reporting then compares two different questions and treats the answers as one number.
Can corrosion rates be rolled up to a corporate integrity dashboard?
Only if the basis travels with the number. A usable roll-up carries the governing rate, which of the two rates governed, the baseline type, the minimum-thickness basis, the date of the last reading and whether an override was applied. Without those fields, a group ranking of shortest remaining life mostly ranks the conservatism of local procedures rather than the actual condition of the equipment.
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.