How API 653 Sets External, Internal and Settlement Frequencies on Atmospheric Tanks
API 653 governs in-service inspection of atmospheric storage tanks built to API 650 or API 12C. It sets routine in-service checks at intervals not exceeding one month, external inspection by an authorized inspector at five years or RCA/4N years — whichever is less — and internal inspection intervals capped at 20 years without a release prevention barrier.
API 653 is the in-service code. API 650 governs how an atmospheric welded steel tank is built; API 653 governs it from the day it enters service, covering the foundation, bottom, shell, structure, roof, appurtenances and nozzles to the face of the first flange. Section 6 sets four separate clocks that run in parallel rather than one master interval: a monthly routine visual by owner personnel, a five-year external inspection by an authorized inspector, an external ultrasonic thickness programme on its own corrosion-rate schedule, and an internal inspection whose interval is built from bottom thickness, corrosion rate and installed safeguards. Section 1.3 subordinates all of it to jurisdictional regulation where the two conflict, and section 6.2.3 acknowledges that jurisdictional regulations control inspection frequency where they apply. PHMSA and EPA both lean on API 653 rather than writing their own intervals.
Source: API Standard 653, Tank Inspection, Repair, Alteration, and Reconstruction, Fifth Edition, November 2014 — sections 1.1, 1.3, 3, 4.4.5, 6.2, 6.3, 6.4, 12.5 and Annex B, with Tables 4.4a/4.4b and 6.1; API 653 Aboveground Storage Tank Inspector Body of Knowledge, March/July/November 2026; 49 CFR 195.432; 40 CFR 112.8(c)(6).
| Inspection | Who performs it | Interval | Clause | What shortens or extends it |
|---|---|---|---|---|
| Routine in-service visual | Owner/operator personnel; need not be an authorized inspector | Consistent with site conditions, not to exceed one month | 6.3.1.1–6.3.1.3 | Findings on leaks, distortion, settlement or foundation escalate to an authorized inspector |
| External visual inspection | Authorized inspector; tank may stay in operation | At least every five years, or RCA/4N years, whichever is less | 6.3.2.1 | A rising shell corrosion rate N shortens it directly; insulation removed only as needed |
| External UT of shell, corrosion rate unknown | Extent determined by owner/operator | Maximum five years | 6.3.3.2 a) | Establishing a rate from tanks in similar service on a five-year measurement interval |
| External UT of shell, corrosion rate known | Extent determined by owner/operator | Smaller of RCA/2N years or 15 years | 6.3.3.2 b) | An out-of-service internal inspection at an equal or shorter interval may substitute |
| Cathodic protection survey | Per API RP 651; owner/operator reviews results | Per API RP 651 | 6.3.4.1 | Owner/operator must assure competency of personnel performing surveys |
| Initial internal inspection | Authorized inspector | 10 years plus cumulative Table 6.1 credits; capped at 20 years, or 30 with a release prevention barrier | 6.4.2.1.1 | Reinforced lining +5, unreinforced lining +2, cathodic protection +5, RPB +10, thicker bottom, stainless bottom +10 |
| Subsequent internal, corrosion-rate method | Authorized inspector | Set by measured bottom corrosion rate and minimum remaining thickness; capped at 20 years, or 30 with an RPB | 6.4.2.2.1 | MRT projected below Table 4.4 forces lining, repair, replacement or a shorter interval |
| Subsequent internal, risk-based | RBI team competent in API RP 580 | Set by the RBI assessment; assessment reviewed at intervals not exceeding 10 years | 6.4.2.2.2 | Corrosion rates from low inspection effectiveness such as spot UT are not permitted |
Which code governs, and which regulators borrow it
API 653 covers steel storage tanks built to API 650 and its predecessor API 12C, and it provides minimum requirements for maintaining their integrity after they have been placed in service. Its scope is limited to the tank foundation, bottom, shell, structure, roof, attached appurtenances, and nozzles to the face of the first flange, first threaded joint or first welding-end connection. Where API 653 and API 650 appear to conflict on an in-service tank, API 653 governs.
The jurisdiction question has an explicit answer in the standard. Section 1.3 states that where any provision presents a direct or implied conflict with a statutory regulation, the regulation governs — but where API 653 is more stringent than the regulation, API 653 governs. Section 6.2.3 reinforces the point, noting that jurisdictional regulations control the frequency and interval of inspections where they apply, and that knowledge of those regulations is necessary for compliant scheduling.
In the United States the two regulators that matter most both borrow rather than write. PHMSA's 49 CFR 195.432 requires operators to inspect in-service breakout tanks according to API 653, with the alternative risk-based internal inspection interval provision carved out. EPA's SPCC rule at 40 CFR 112.8(c)(6) requires integrity testing to industry standards without prescribing a frequency, and requires visual inspection to be combined with another technique such as hydrostatic, radiographic, ultrasonic or acoustic emission testing.
Routine in-service inspection: the one-month clock
The most frequent requirement in API 653 is also the one most often missed in programme audits. Clause 6.3.1.1 requires that the external condition of the tank be monitored by close visual inspection from the ground on a routine basis. This inspection may be done by owner/operator personnel and by people who are not authorized inspectors, provided they are knowledgeable about the storage facility operations, the tank, and the characteristics of the product stored.
Clause 6.3.1.2 sets the interval: consistent with conditions at the particular site, but not to exceed one month. That is a hard ceiling, not a target. A tank on a quarterly walk-down schedule is non-compliant regardless of how thorough the quarterly walk-down is, and no corrosion argument extends it.
Clause 6.3.1.3 defines the content. The routine inspection covers the tank's exterior surfaces, and evidence of leaks, shell distortions, signs of settlement, corrosion, and the condition of the foundation, paint coatings, insulation systems and appurtenances should be documented for follow-up by an authorized inspector. The monthly walk is the detection layer that decides whether the five-year external inspection finds a surprise. Programmes that struggle to evidence twelve records a year per tank usually have a records problem rather than an inspection problem, which is what our NDT program audit and gap assessment is built to surface.
The five-year external inspection and the RCA/4N test
Clause 6.3.2.1 requires that all tanks be given a visual external inspection by an authorized inspector, conducted at least every five years or RCA/4N years, whichever is less. RCA is the remaining corrosion allowance — the difference between the measured shell thickness and the minimum required thickness, expressed in mils. N is the shell corrosion rate in mils per year. Tanks may be in operation during this inspection.
The formula is the part that gets skipped. Five years is a ceiling, not a default, and a tank with a thin shell and an active corrosion rate falls below it arithmetically. A shell with 40 mils of remaining corrosion allowance corroding at 4 mils per year yields RCA/4N = 40/16 = 2.5 years, and the external inspection is due at 2.5 years, not five. Every external interval decision should show that calculation on the face of the report.
Two supporting requirements travel with the external inspection. Clause 6.3.2.2 permits insulation to be removed only to the extent necessary to determine the condition of the exterior wall or roof, which keeps insulated-tank inspections proportionate. Clause 6.3.2.3 requires that grounding system components such as shunts and mechanical cable connections be visually checked, referencing API 2003 for hydrocarbon ignition prevention.
External UT of the shell: five years, or RCA/2N to a 15-year ceiling
Clause 6.3.3.1 describes external ultrasonic thickness measurement of the shell as a means of determining a rate of uniform general corrosion while the tank is in service, with the extent of measurement determined by the owner/operator. It is optional in the sense that the standard says when used — but once a programme relies on it, clause 6.3.3.2 fixes the intervals.
Where the corrosion rate is not known, the maximum interval is five years, and the standard permits rates to be estimated from tanks in similar service based on thickness measurements taken at an interval not exceeding five years. Where the corrosion rate is known, the maximum interval is the smaller of RCA/2N years or 15 years. Note the difference from the external visual: the UT formula divides by 2N, the visual by 4N, so the visual always falls due first on the same tank data.
Clause 6.3.3.3 provides the substitution most owners want. Internal inspection of the tank shell, performed when the tank is out of service, can be substituted for a programme of external ultrasonic thickness measurement if the internal inspection interval is equal to or less than the interval that 6.3.3.2 b) would require. That is the clause that lets a tank on a tight internal cycle avoid running a parallel external UT programme.
The initial internal interval: 10 years plus Table 6.1 credits
Clause 6.4.2.1.1 sets the interval from initial service date to first internal inspection at not more than 10 years, unless the tank carries one or more of the leak prevention, detection, corrosion mitigation or containment safeguards listed in Table 6.1. The maximum initial interval is 10 years plus incremental credits for those safeguards, and the credits are cumulative.
Table 6.1 prices each safeguard. A thick-film reinforced lining on the product side installed per API RP 652 adds five years. A thin-film or thick-film unreinforced lining per API RP 652 adds two. Cathodic protection of the soil side installed, maintained and inspected per API RP 651 adds five. A release prevention barrier installed per API 650 Annex I adds ten. A qualifying stainless steel bottom per API 650 Annex SC with Annex S or X adds ten. Initial bottom thickness above 6.0 mm (0.25 in.) adds a calculated credit using 0.375 mm per year (15 mpy) or a rate from Annex H.
The standard's own worked example makes the arithmetic concrete: an 8 mm bottom with a release prevention barrier and a fibreglass-reinforced lining earns 10 years initial, plus 5 for the lining, plus 10 for the RPB, plus 4.2 for the thickness credit, giving 29.2 years. Two ceilings bind that result — 20 years without a release prevention barrier, 30 years with one. Clause 6.4.2.1.1 also states that the limits of Table 4.4 do not apply when establishing the initial interval this way.
Subsequent internal intervals: minimum remaining thickness drives the date
Clause 6.4.2.2.1 sets the subsequent interval using the measured tank bottom corrosion rate and the minimum remaining thickness in accordance with 4.4.5, with a maximum of 20 years without a release prevention barrier and 30 years with one. The interval is an output of the thickness calculation, not an input chosen first and justified afterwards.
The governing equation in 4.4.5.1 is MRT = (minimum of RTbc or RTip) − Or(StPr + UPr), where Or is the in-service interval of operation in years to the next internal inspection, RTbc and RTip are minimum remaining thicknesses from bottom-side and internal corrosion after repairs, and StPr and UPr are the top-side and bottom-side corrosion rates not repaired. StPr is zero for coated areas where the coating life equals or exceeds Or; UPr is zero for areas with effective cathodic protection.
The result is tested against Table 4.4. In USC units the minimum bottom plate thickness at next inspection is 0.10 in. for a tank bottom and foundation design with no means for detection and containment of a bottom leak, and 0.05 in. where such means exist or where a reinforced lining thicker than 0.05 in. is applied per API 652. Clause 4.4.5.3 states the consequence directly: if the projected thickness falls short, the bottom is lined, repaired or replaced, or the interval is shortened. In the critical zone, 4.4.5.4 sets a separate floor and prohibits thickness averaging.
The risk-based route and its guardrails
Clause 6.4.2.2.2 permits an owner/operator to set the subsequent internal interval using risk-based inspection procedures in accordance with API RP 580. The assessment must be a systematic evaluation of both the likelihood and the consequence of failure, thoroughly documented, and it must produce an inspection strategy defining methods, frequencies for internal, external and in-service inspections, and mitigation steps.
The guardrails are stricter than most programmes expect. The assessment must be performed by a team with inspection and engineering expertise in API RP 580 principles, tank design, construction and deterioration modes, and that team must review and approve it at intervals not exceeding 10 years, or sooner where process, equipment or consequence changes warrant. The methodology itself requires a documented validation review, and API recommends that validation be performed by an entity external to the assessment team.
One requirement disqualifies a great deal of existing data: corrosion rates derived from prior inspections must come from high or medium inspection effectiveness as defined in the owner's procedures, with API RP 581 given as the reference for those categories. Rates from low inspection effectiveness, spot UT explicitly named, must not be used.
Settlement surveys: how many points, and where
The point count comes from clause 12.5.2. The tank is surveyed with an even number of elevation measurement points, N, uniformly distributed around the circumference, with the minimum given by N = D/10 where D is the tank diameter in feet. All values of N round up to the next higher even whole number, N is never fewer than eight, and the maximum spacing between measurement points is 10 m (32 ft). Annex B Figures B.1 and B.2 reference that method and add that there must be at least four equally spaced diametrical measurement lines.
Clause 12.5.1 requires a settlement survey for all existing tanks undergoing a hydrostatic test, except tanks with a documented service history of acceptable settlement values where no settlement is anticipated during the hydrotest. An initial survey with the tank empty provides the baseline; in its absence, the standard directs that the tank be assumed initially level, which is a conservative and often expensive assumption.
Annex B adds measurement discipline that changes results. Where bottom distortion or corrosion extends beyond the shell, readings taken near bottom lap welds can produce significant elevation errors, and surveying the weld between the first and second shell courses gives more consistent results. Where measured settlement approaches the allowable, Annex B directs repeating the measurement with the bottom forced into contact with the soil, because a bottom not bearing on the foundation over- or under-states settlement.
Acceptance criteria for differential settlement on a ringwall
Annex B separates settlement into three components: uniform settlement, rigid body tilt in a plane, and out-of-plane settlement. Only the third threatens shell and bottom integrity, so the evaluation subtracts the first two. The uniform component is the lowest point on the plotted curve; the rigid tilt plane is represented by an optimum cosine curve of the form Elevpred = a + b·cos(θ + c), fitted by least squares.
The validity test is explicit and is where most evaluations fail quietly. The optimum cosine curve is only considered valid — meaning it accurately fits the measured data — if R² is greater than or equal to 0.9. Where out-of-plane settlement concentrates in one or two areas, the least-squares fit under-predicts the local settlement and is not conservative, and R² typically falls below 0.9. That result routes the tank to the B.2.2.5 and B.3.2.2 procedure, which uses a K factor indexed by tank diameter and roof type, or to a more rigorous engineering assessment.
Where the cosine fit is valid, permissible out-of-plane deflection follows Smax = 11L²Y/(2EH). In USC units Smax and L and H are in feet, with L the arc length between measurement points, Y the shell yield strength in psi, E Young's modulus in psi and H the tank height. Annex B credits the equation to Marr, Ramos and Lambe, Criteria for Settlement of Tanks, ASCE Journal of the Geotechnical Engineering Division, Volume 108, August 1982.
Bottom depressions, edge settlement and what triggers NDE
Annex B contains a trap worth knowing before the survey is commissioned. Taking more measurement points improves the cosine fit but shrinks the arc length L, and because Smax scales with L², using every point in the Smax equation produces a very small allowable settlement. The standard's resolution is to compute the optimum cosine curve using all measurements, then calculate Si and Smax from a subset spaced no further apart than 10 m (32 ft), minimum eight points, with the subset required to include the points furthest from the optimum cosine curve.
For depressions and bulges away from the shell, clause B.3.3 gives the permissible value as BB = 0.37R in USC units, where BB is the maximum bulge height or depression depth in inches and R is the radius of the inscribed circle in the affected area in feet. In SI units the relationship is BB = 0.031R with both terms in millimetres. Those limits apply to bottoms with single-pass lap-welded joints.
Edge settlement is evaluated against allowable values Bew and Be read from Figures B.11 and B.12, with Bew the more conservative and the sensible first screen for all settled areas. Locating the breakover point takes a straight edge laid on the unsettled bottom, and on a cone-up or cone-down bottom the settlement B is measured from a projection of the unsettled bottom rather than from level. Clause B.3.4.2 sets the NDE trigger: where measured settlement exceeds 75 % of Bew, welds within 300 mm (12 in.) of either side of the breakover area are examined visually, and suspect areas by magnetic particle or liquid penetrant examination.
Turning the clauses into a programme that survives an audit
Every interval above is defensible only if the evidence behind it is retrievable — the thickness data feeding RCA and N, the Table 6.1 credits claimed and the installation records supporting them, the MRT calculation with its assumed Or, the settlement point count against N = D/10, and the certification status of the inspector who signed. Programmes fail audits on retrieval far more often than on technical judgement.
Two structural weaknesses recur. The first is inspector certification lapsing mid-cycle, which puts a signature on a report that the code does not recognise; tracking expiry alongside CPD evidence is what our NDT personnel certification tracking module in the Atlantis ERP exists to automate. The second is a monthly routine inspection performed diligently and recorded informally, which produces a compliant tank and a non-compliant record set.
Where the interval calculations themselves need independent review, or where a tank has failed its thickness or settlement criteria and the question becomes whether it can stay in service, our tank inspection services provide NDT inspectors and the inspection data that the owner's fitness-for-service evaluation under API 579 then relies on. For sites where the tank programme sits inside a wider PSM obligation, our OSHA PSM mechanical integrity support ties the inspection evidence to the mechanical integrity element. Affordable, accessible and fully customizable, with a demo or quote on request via contact.
What code governs external inspection intervals for atmospheric petroleum storage tanks?
API 653 does, as the in-service code for tanks built to API 650 or its predecessor API 12C. It sets the external inspection at five years or RCA/4N years, whichever is less, performed by an authorized inspector with the tank in operation. Section 1.3 gives precedence to any conflicting statutory regulation, and where API 653 is more stringent than the regulation, API 653 governs.
How do federal regulations use API 653?
They incorporate it rather than replacing it. PHMSA's rule at 49 CFR 195.432 requires operators to inspect in-service breakout tanks according to API 653, while excluding its alternative risk-based internal inspection interval provision — so a regulated breakout tank cannot use RBI to extend the internal interval. EPA's SPCC rule at 40 CFR 112.8(c)(6) is performance-based and points to industry standards including API 653.
What sets the out-of-service internal inspection interval?
Two permitted routes. The corrosion-rate method uses the measured tank bottom corrosion rate and the minimum remaining thickness calculated per 4.4.5, capped at 20 years without a release prevention barrier and 30 years with one. The risk-based method under 6.4.2.2.2 uses an API RP 580 assessment of likelihood and consequence, reviewed at intervals not exceeding 10 years.
What extends an initial internal inspection interval past 10 years?
Table 6.1 safeguards, and they are cumulative. A thick-film reinforced product-side lining per API RP 652 adds five years, an unreinforced thin or thick film adds two, soil-side cathodic protection per API RP 651 adds five, a release prevention barrier per API 650 Annex I adds ten, and a qualifying stainless bottom adds ten. Initial bottom thickness above 6.0 mm adds a calculated credit.
How many elevation points does a settlement survey need?
API 653 gives the minimum as N = D/10 with D in feet, rounded up to the next higher even whole number and never fewer than eight points, uniformly distributed around the circumference. Maximum spacing between points is 10 m (32 ft). Annex B Figure B.1 adds that there must be at least four equally spaced diametrical measurement lines.
What is the acceptance criterion for out-of-plane settlement?
Where a rigid tilt plane fits the data, permissible out-of-plane deflection is Smax = 11L²Y/(2EH), with L the arc length between measurement points, Y the shell yield strength, E Young's modulus and H the tank height. The optimum cosine curve is only valid when R² is 0.9 or greater. Exceedance routes the tank to B.3.2.2 or a rigorous engineering assessment.
The out-of-service internal inspection interval in one answer
API 653 sets the first out-of-service internal inspection at no more than 10 years after a tank enters service, then lets the owner add cumulative credits for safeguards such as a product-side lining, soil-side cathodic protection, a release prevention barrier (RPB), a qualifying stainless bottom or extra original bottom thickness. The initial interval cannot exceed 20 years without an RPB or 30 years with one. After that first inspection, the next interval comes from the measured bottom corrosion rate and the minimum remaining thickness, under the same 20-year and 30-year caps, or from an owner-commissioned risk-based assessment. Jurisdictional rules override all of it where they are stricter.
The sections above walk through the clause mechanics. This section adds what the clause text leaves to the people running the programme: the factors API 653 says can shorten or lengthen an interval, the evidence an authorized inspector expects before accepting a credit, worked scenarios for common tank configurations, the in-service alternative for bottom-only assessment, the NDE that actually produces the bottom data, and the regulatory overlays in the United States and Canada. Clause references are to the Fifth Edition. Addendum 4 (July 2025) revised the text of the initial-interval clause and the requirements for tank-bottom plate thickness measurement, so confirm every number against the edition and addendum your programme is written to before relying on it.
What API 653 says can shorten or lengthen the interval
Section 6.2 opens the inspection-frequency discussion with a list of considerations rather than a single number. Several factors determine how often a tank should be inspected, and the standard expects the owner and the authorized inspector to weigh them for each tank rather than copy one interval across a tank farm. The 10-year baseline and the Table 6.1 credits are ceilings that apply only when those considerations do not point to something shorter.
The factors fall into five practical groups:
- The product and how it attacks steel. The nature of the product stored, its temperature, water and sediment content, and whether it contains corrosive species all change the product-side rate. Crude with settled water and sulfate-reducing bacteria behaves very differently from a dry finished product. A change of service, for example converting a gasoline tank to biodiesel or to a heated product, resets the assumptions behind the previous interval.
- Corrosion rates and allowances. Measured rates from previous inspections, the remaining corrosion allowance and the reliability of the data behind those numbers are the strongest single input. A tank whose last inspection found isolated deep pitting cannot carry forward an interval that assumed uniform thinning.
- Corrosion prevention systems. Linings, coatings and cathodic protection only earn their value if they are installed to the relevant recommended practice and kept in service. A lining past its expected life or a CP system with lapsed survey records no longer supports a long interval.
- Consequence and detection. The location of the tank, including its proximity to water, groundwater sensitivity, populated areas and secondary containment, sits alongside leak detection capability and the presence of a double bottom or RPB. A tank that would release into a waterway if the bottom failed warrants a shorter, more conservative cycle than an identical tank inside an impermeable dike with leak detection.
- Operation and regulation. Changes in operating mode, such as frequent fill and draw cycling, heated service or a move from storage to rapid transfer duty, can accelerate bottom and shell damage. Jurisdictional requirements are listed explicitly; where a regulator prescribes a shorter interval, it governs.
The practical consequence is that the interval on an inspection report should carry its reasoning: the baseline, each credit claimed, the factors considered and any reason the inspector chose something shorter than the maximum. An interval written as a bare date is hard to defend at audit and impossible to re-evaluate when service conditions change.
Evidence an authorized inspector expects before accepting a Table 6.1 credit
Credits are where most extended initial intervals fail review. Table 6.1 does not award time for the existence of a safeguard; it awards time for a safeguard installed and maintained to a named recommended practice. The authorized inspector setting the interval needs records, not assurances. The typical evidence for each credit looks like this:
| Safeguard claimed | Reference practice named by API 653 | Records an inspector will typically ask for | Common reason the credit is refused |
|---|---|---|---|
| Product-side bottom lining (reinforced or unreinforced) | API RP 652 | Lining specification, surface preparation records, coating inspection and holiday test results, applicator documentation, date applied | No installation records; lining applied over unknown surface condition; lining type cannot be confirmed as reinforced |
| Soil-side cathodic protection | API RP 651 | CP design, commissioning survey, periodic survey results showing protection criteria met, rectifier records | System installed but surveys missing or showing inadequate protection; CP added years after service start without baseline data |
| Release prevention barrier | API 650 Annex I | Design drawings of the double bottom or liner, leak detection arrangement, construction records | Barrier exists but leak detection path is blocked or undocumented; barrier not designed to the referenced annex |
| Stainless steel bottom | API 650 annexes referenced in Table 6.1 | Material certificates, fabrication records confirming the annex requirements | Material or construction records unavailable |
| Bottom thicker than the base value | Table 6.1 thickness credit | Original mill or construction records showing as-built bottom thickness | As-built thickness unknown, so the credit cannot be calculated |
Two points catch owners out repeatedly. First, the credits are tied to the condition that makes them valid: a CP credit implies the system has protected the bottom since service began, so a system added late or left unsurveyed for years undermines the basis. Second, missing construction records are common on older tanks and acquisitions. Where the as-built bottom thickness or lining history cannot be shown, the conservative course is to take the 10-year baseline, inspect, and let measured data drive the subsequent interval.
Worked scenarios for common tank configurations
The scenarios below apply the Fifth Edition credit values described earlier on this page. They are illustrations of the arithmetic, not interval recommendations; the interval for a real tank is set by the owner and the authorized inspector after considering every factor in Section 6.2 and the edition and addendum in force.
Scenario 1: bare carbon steel bottom, no CP, no lining, no RPB, base bottom thickness. No credits apply. The initial internal inspection falls due at no more than 10 years after the tank entered service. This is the default many older terminal tanks sit on, and it is also the right answer whenever credit evidence is missing.
Scenario 2: carbon steel bottom with soil-side CP per API RP 651 and a thin-film product-side lining per API RP 652, no RPB. The arithmetic gives 10 plus 5 for CP plus 2 for the unreinforced lining, or 17 years. That is below the 20-year cap that applies without an RPB, so the arithmetic result stands as the maximum, subject to the Section 6.2 factors.
Scenario 3: same as scenario 2 but with a reinforced thick-film lining and a bottom 3/8 in. thick. The lining now earns 5 rather than 2, giving 10 plus 5 plus 5, or 20, before the thickness credit. The thickness credit adds the excess over 0.25 in. divided by the default rate of 15 mils per year: 0.125 in. divided by 0.015 in. per year is roughly 8.3 years. The sum exceeds 28 years, but without an RPB the initial interval is capped at 20. The extra credits do not carry forward; they simply show that the cap, not the calculation, is governing.
Scenario 4: new tank with an RPB per API 650 Annex I, CP and a reinforced lining. 10 plus 10 plus 5 plus 5 gives 30 years, which meets the 30-year cap that applies with an RPB. Any thickness credit would be absorbed by the cap.
The lesson from scenarios 3 and 4 is that the RPB is what moves the ceiling. Owners planning new construction or a bottom replacement who want a long first cycle should weigh the barrier design at that stage, because it is the only safeguard that raises the cap rather than adding time beneath it.
The in-service alternative when only the bottom needs assessment
API 653 recognises that taking a tank out of service, cleaning it and gas-freeing it is costly and carries its own risk. Where the purpose of the internal inspection is solely to determine the condition and integrity of the tank bottom, the Fifth Edition permits the assessment to be carried out with the tank in service using on-stream methods. These include robotic ultrasonic thickness measurement and other in-service inspection techniques, which may be supplemented by electromagnetic methods, provided the data gathered are sufficient for the authorized inspector to establish the next interval.
The conditions matter more than the permission. The data must be good enough to set an interval, which means adequate coverage of the bottom, including the critical zone near the shell, and enough confidence in measured remaining thickness and corrosion rate. In practice, in-service robotic surveys work best where the product is compatible with the crawler, visibility or coupling conditions are acceptable, and sludge or debris will not mask large areas of plate. They do not inspect the shell interior, roof structure, internal piping or welds in the way an out-of-service visual can, so they answer the bottom question only. Where an in-service survey leaves gaps or ambiguous indications, the authorized inspector may still require an out-of-service inspection.
NDE methods that produce the bottom data, and their limits
The interval is only as good as the bottom data behind it. An out-of-service internal inspection normally combines a screening method for full-floor coverage with a quantitative method for sizing. The table below summarises what each method contributes.
| Method | What it finds | Limits to plan around |
|---|---|---|
| Magnetic flux leakage (MFL) floor scanning | Rapid screening of plate for topside and soil-side metal loss across most of the floor | Screening tool; indications need sizing by UT; reduced sensitivity near shell, lap welds and obstructions; thick coatings and plate thickness affect detection |
| Ultrasonic thickness (UT) prove-up | Remaining thickness at MFL indications and at grid or critical-zone locations | Point measurements; coverage depends on the plan; soil-side pitting can be missed between points |
| UT corrosion mapping or phased array | Area thickness maps for annular plates, critical zone and suspect regions | Slower than screening; surface condition and access affect results |
| Visual examination | Pitting, coating failure, weld condition, settlement distortion, internal component damage | Needs cleaned surfaces and adequate lighting; cannot see soil-side loss |
| Vacuum box testing | Through-wall leaks in bottom welds and repairs | Leak test only, not a thickness measurement |
| Magnetic particle or liquid penetrant | Surface cracking at welds, especially shell-to-bottom and near settlement breakover areas | Surface methods; need preparation |
Two programme decisions follow. Coverage drives confidence: a floor scanned edge to edge with MFL and proven up with UT produces a corrosion rate an inspector can stand behind, while a handful of spot readings does not. If the owner later wants to use a risk-based route for the subsequent interval, the Fifth Edition does not allow corrosion rates from low-effectiveness inspection such as spot UT to be used, so the coverage choice at this inspection shapes the options at the next one. Second, Addendum 4 revised the tank-bottom thickness measurement requirements and points to API RP 575; confirm the measurement approach in your inspection plan against the current addendum before the tank is opened.
Regulatory overlays in the United States and Canada
API 653 is a consensus standard. It becomes mandatory through regulation, contract or the owner's own programme, and several regulators adopt it with their own conditions.
- PHMSA breakout tanks. 49 CFR 195.432 requires operators to inspect in-service atmospheric and low-pressure steel breakout tanks in accordance with API 653, but excludes the standard's risk-based alternative internal inspection interval provision. A regulated breakout tank therefore relies on the corrosion-rate route for subsequent intervals.
- EPA SPCC. 40 CFR 112.8(c)(6) requires bulk storage containers to be tested or inspected for integrity on a regular schedule according to industry standards, combining visual inspection with another technique. API 653 is the usual standard for field-erected tanks; STI SP001 is often used for shop-built tanks. The SPCC plan, certified by a Professional Engineer where required, should state which standard and interval apply to each container.
- OSHA PSM. Where tanks fall inside a covered process, 29 CFR 1910.119(j) requires inspection and testing that follows recognized and generally accepted good engineering practices, at a frequency consistent with those practices and with prior operating experience. API 653 is commonly cited as that practice for atmospheric tanks.
- States and local authorities. State environmental agencies, fire marshals and local codes can impose their own tank inspection or leak detection requirements. Confirm with the authority having jurisdiction.
- Canada. Requirements are provincial and territorial, typically through fire codes and environmental regulation of storage tank systems, and federally regulated sites have their own rules. Many Canadian owners use API 653 as the inspection basis; confirm the applicable provincial regulation and any prescribed frequency with the regulator.
Common mistakes that shorten intervals or fail audits
- Claiming credits without records. A lining or CP credit with no installation or survey evidence will not survive review, and an interval built on it becomes non-compliant retroactively.
- Applying a fleet interval. Copying one interval across tanks in different services ignores Section 6.2 and is a common audit finding.
- Treating the cap as the target. 20 and 30 years are maximums. The calculated interval or the inspector's judgement often governs at a shorter date.
- Thin data at the first inspection. Limited floor coverage produces an uncertain corrosion rate, which forces a shorter subsequent interval and can rule out other routes later.
- Ignoring service changes. A new product, heating or higher cycling invalidates the assumptions behind the existing interval until it is re-evaluated.
- Using an outdated edition. Programmes written to an earlier edition or addendum may cite clause numbers and text that have since changed.
How Atlantis supports out-of-service and in-service tank bottom inspections
Atlantis performs the NDE that produces the bottom data: MFL floor scanning, UT prove-up and critical-zone thickness measurement, corrosion mapping, visual examination, vacuum box testing and MT or PT at welds, carried out by ASNT-certified technicians under ASNT Level III oversight. Results are delivered to the owner's API 653 authorized inspector, who remains the inspector of record and sets the next interval. Atlantis does not set intervals, perform risk-based or fitness-for-service assessments, or act as the authorized inspector. See our aboveground storage tank inspection and MFL testing pages, or request a quote within 24 hours.
More questions on the API 653 internal inspection interval
What does API Std 653 require for out-of-service internal inspection intervals on aboveground welded steel storage tanks?
A first internal inspection no later than 10 years after service begins, extendable by cumulative safeguard credits up to 20 years without an RPB or 30 years with one. Subsequent intervals come from measured bottom corrosion rates and minimum remaining thickness under the same caps, or from an owner-commissioned risk-based assessment where the jurisdiction permits it.
What factors shorten the API 653 internal inspection interval?
Aggressive product or contaminants, high measured corrosion rates, unreliable or limited inspection data, failed or undocumented linings and CP, sensitive locations, weak leak detection, increased fill cycling or heated service, a change of service, and jurisdictional rules that prescribe shorter intervals.
Does a release prevention barrier extend the interval?
Yes. In the Fifth Edition an RPB per API 650 Annex I earns a 10-year credit and raises the cap from 20 to 30 years, for both the initial and subsequent intervals.
Can a tank bottom be inspected without taking the tank out of service?
Where the internal inspection is solely to assess the bottom, API 653 permits in-service methods such as robotic ultrasonic thickness measurement, which may be supplemented by electromagnetic methods, if the data are sufficient to set the next interval.
Can PHMSA-regulated breakout tanks use risk-based intervals?
No. 49 CFR 195.432 adopts API 653 for breakout tanks but excludes its risk-based alternative internal inspection interval provision.
Who sets the internal inspection interval?
The owner/operator with its API 653 authorized inspector, within the code and any jurisdictional requirement. NDE contractors provide the data but do not set the interval.
Does the first inspection establish the corrosion rate?
It establishes the first measured bottom rate, which then drives the subsequent interval through the minimum remaining thickness calculation. Better coverage gives a more defensible rate.
What changed in API 653 Addendum 4?
Addendum 4, published July 2025, revised the initial internal interval clause and the requirements for tank-bottom plate thickness measurement, with reference to API RP 575. Check the full addendum before updating inspection plans.
Planning a tank outage or an in-service floor survey? Talk to our inspection team, compare MFL and UT floor scanning, read about robotic tank-bottom inspection, review the API 653 audit evidence pack, or get a floor-scan quote. Related: API 653 tank inspection guide and corrosion mapping.