Aboveground storage tanks run on several inspection clocks at once, and API 653 sets each one differently. The external clock is calendar-driven with a corrosion-rate override. The internal clock is driven almost entirely by one question: how fast is the bottom corroding, and how long until it reaches minimum thickness. Programs that treat these as one generic interval end up either over-inspecting shells or under-inspecting bottoms.
The external clocks
Routine in-service checks are frequent and light: someone walks the tank, looks at the shell, the foundation, the appurtenances, and records that they did. The formal external inspection is less frequent and belongs to an authorized inspector, on a cycle set by the calendar or by a fraction of the shell's remaining corrosion life, whichever comes first.
Shell thickness surveys run on their own schedule. When the corrosion rate is known, the interval derives from remaining life on the shell courses. When it is not known, the standard forces a shorter default until data exists. Either way, the record has to show which case applied and why.
The internal clock and the bottom
The internal inspection interval is where API 653 concentrates its attention, because the tank bottom is the component you cannot monitor from outside in any complete way. The logic is projection: take the minimum remaining thickness on the bottom, take the corrosion rates on both sides of the plate, product side and soil side, and compute when the bottom reaches its minimum acceptable thickness. The next internal inspection has to happen before that date.
Several things follow from that logic:
- The soil-side rate matters as much as the product side. It is also the harder one to know, which is why bottom condition at the last internal, cathodic protection status, and release prevention barriers all feed the interval decision.
- A similar-service argument needs documentation. Where rates are inferred from comparable tanks rather than measured, the comparison basis belongs in the file.
- Risk-based extension is available but not free. An RBI assessment can push the internal interval past the default maximum, but the assessment itself becomes a record the auditor will read, including when it was last revalidated.
What the tank record must reconstruct
An auditor working a tank file wants to rebuild each clock from evidence:
- The last internal inspection report, with bottom plate thickness data, patch and repair locations, and the inspector's evaluation.
- The corrosion rates used, with their source: measured, similar service, or conservative default.
- The projection math, from remaining thickness to the next internal date.
- The external and UT survey history, showing the calendar cycles were kept or justifiably adjusted.
- Repairs and alterations, with their examination records tied to the same tank.
The common failure is that the projection lives in a spreadsheet owned by one engineer, while the thickness data lives in survey reports and the repair history lives in the maintenance system. Each piece is fine. The chain between them is what the audit tests, and chains that live in one person's head do not survive personnel changes.
Setting the interval you can defend
The defensible position is always the same shape: here is the thickness, here is the rate and where it came from, here is the projection, here is the date, and here is who signed it. When any of those elements is missing, the interval defaults to conservative, and the tank comes out of service earlier than the metal required.
intelliSPEC keeps the bottom data, the rates, and the projection in one record that assembles as the inspection happens. If you have a tank internal coming up, bring its file to a demo and see the chain built end to end.
