How API 653 Sets External, Internal and Settlement Frequencies on Atmospheric Tanks

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. Programmes building this capability from scratch should start with our RBI program design service rather than retrofitting a spreadsheet, and remember that PHMSA excludes this route for regulated breakout tanks.

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 API 653 tank inspector services provide certified inspectors and our fitness-for-service assessment under API 579 handles the evaluation that follows. 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.

Request a consultation