Is the maximum tank internal inspection interval 20 years, or 30?
Ask around a tank farm what the longest permitted gap between internal inspections is, and you will hear both 20 years and 30 years with equal confidence. Both are right. What decides which applies is whether the tank has a release prevention barrier.
The two ceilings
API 653 caps the internal inspection interval at:
- 20 years for a tank with no release prevention barrier
- 30 years for a tank that has one
The reasoning is about consequence rather than deterioration. The bottom of a tank corrodes at whatever rate its service and its soil dictate, and a barrier does not slow that down by a single millimetre. What a barrier changes is what happens when the floor finally lets go. Without one, product goes into the ground, unseen, possibly for a long time. With a barrier and a means of detection beneath it, the same failure is contained and announces itself.
The same logic runs through the minimum bottom thickness rules, which is a useful cross-check that you have understood it rather than memorised it:
| Bottom arrangement | Minimum thickness at next inspection |
|---|---|
| No means of detecting and containing a leak | 2.54 mm (0.10 in.) |
| Detection and containment provided | 1.27 mm (0.05 in.) |
| Reinforced lining over 1.27 mm to API 652 | 1.27 mm (0.05 in.) |
Halve the consequence and the code halves the metal it insists on finding.
Both figures are ceilings, not entitlements
This is the part that causes real trouble on site. Neither 20 nor 30 years is an interval you are granted. They are limits that sit above the interval you calculate.
The calculated interval comes from how much bottom is left and how fast it is going — the minimum remaining thickness at the end of the period, worked from the bottom-side and product-side corrosion rates acting together. That figure very often lands well short of the ceiling. A tank in sour crude with an actively corroding floor might calculate out at nine years, and no barrier will stretch it to thirty.
So the sequence is always: calculate first, then apply whichever ceiling the tank’s arrangement earns, and take the lower of the two.
Two details worth carrying
Effective cathodic protection lets the underside corrosion rate be taken as zero in that calculation. That is a large concession, and it is why proving the system actually works by survey — structure-to-soil potential, not rectifier output — matters far more than having one installed.
A coating lets the product-side rate be taken as zero only if the coating outlives the interval. A lining expected to fail halfway through cannot justify a zero rate for the whole period, because the steel is bare for the rest of it.
Why this note exists
We had this wrong in our own API 510 and API 653 practice material until recently: a question stated flatly that 20 years was the maximum, with no mention of the barrier. It was corrected once we checked the figure against the code itself rather than against recollection.
That is the general lesson, and it is worth more than the number. Interval questions are where practice material ages worst, because the figures move between editions and addenda. If a source cannot tell you which edition it was written to, it cannot tell you whether its intervals are current.