Minimum Acceptable Shell Thickness for API 653 Storage Tanks

API 653 calculates the minimum acceptable thickness of a welded steel tank shell from the liquid height above the point assessed, the tank diameter, the product's specific gravity, the allowable stress for that course and the joint efficiency. A whole course is checked at one foot above its bottom; a locally thinned area uses the full height above it. No course may fall below 0.1 in.

In U.S. units the whole-course form is t-min = 2.6 (H - 1) D G / (S E), and the locally thinned area form drops the one-foot reduction, with t-min in inches, H and D in feet and S in pounds per square inch. The constant is not arbitrary: it is the hydrostatic pressure of water per foot of depth, multiplied by twelve and halved, so the formula is the hoop-stress equation for a cylinder under liquid head. The allowable stress comes from the code's table for the plate material, with lower values for the bottom two courses, and the joint efficiency is the original weld joint efficiency unless the area is far enough from any weld for 1.0 to apply. The authorized inspector compares t-min with measured thickness, not nominal, to decide whether the course is acceptable or needs repair, rerating or a fitness-for-service assessment.

Source: API Standard 653, Tank Inspection, Repair, Alteration, and Reconstruction, Fifth Edition (November 2014) carried through Addendum 3 (November 2023): Section 4.3 tank shell evaluation, including the minimum thickness calculation in 4.3.3, the evaluation of corroded areas and widely scattered pits, Table 4.1 allowable stresses and Table 4.2 joint efficiencies; Section 6 inspection intervals. API Standard 650 for the one-foot design method the calculation mirrors; API 579-1/ASME FFS-1 for fitness-for-service. Confirm the edition and addenda the owner and jurisdiction invoke.

Technically reviewed by Anoop Rayavarapu — ASNT NDT Level III (UT, RT, MT, PT, VT)
API 653 shell t-min: each term, its source and the usual error
TermMeaningSource in API 653Common error
t-minMinimum acceptable thickness, in.Section 4.3 calculationComparing it with nominal rather than measured thickness
HHeight from the bottom of the course, or of the thinned area, to the maximum liquid level, ftSection 4.3Measuring from the tank bottom for an upper course
DNominal tank diameter, ftSection 4.3Using a surveyed outside diameter
GHighest specific gravity of the contentsSection 4.3Using today's product instead of the heaviest intended
SMaximum allowable stress for the plate material, psiTable 4.1Using the upper-course value on the bottom two courses
EOriginal joint efficiency, or 1.0 away from weldsTable 4.2 and Section 4.3Taking 1.0 for an area on or next to a weld
U.S. customary units, for tanks within the diameter range the one-foot form covers. SI users apply the same relationship with 4.9 in place of 2.6, metres for D and H, megapascals for S and 0.3 m in place of one foot.

Minimum acceptable shell thickness calculation for an aboveground storage tank

API 653 treats the shell as a cylinder holding a column of liquid. The hoop stress at any height depends on the head of liquid above that point, so the thickness the shell needs falls course by course from the bottom up. The calculation gives the thickness at which the plate, at its allowable stress and with its joint efficiency, still holds the maximum liquid level of the heaviest product the tank is allowed to store.

The inspection question is then simple to state: is the relevant measured thickness at or above t-min for that location? If it is, the shell is acceptable for the present service, and the margin above t-min divided by the corrosion rate drives the next inspection date. If it is not, the owner's options are to lower the maximum fill height, change the product, repair the shell, or have the area assessed for fitness for service.

The formula and where each term comes from

In U.S. customary units, t-min = 2.6 (H - 1) D G / (S E) for a full course. H is measured in feet from the bottom of the course being assessed to the maximum liquid level, D is the nominal diameter in feet, G is the highest specific gravity of the contents, S is the maximum allowable stress in pounds per square inch and E is the joint efficiency.

The constant 2.6 is the hydrostatic pressure of water, about 0.433 psi per foot of depth, multiplied by 12 to turn a diameter in feet into inches and divided by 2 from the hoop-stress equation. In SI units the same relationship uses 4.9, with D and H in metres, S in megapascals, t-min in millimetres and 0.3 m in place of one foot. The derivation matters because it shows the formula is pure membrane stress under liquid head; nozzle loads, settlement, wind and seismic loading are separate considerations the shell evaluation has to cover alongside it.

Whole course versus a locally thinned area

The whole-course form subtracts one foot from H because the design method behind it checks each course at a point one foot above its bottom, where restraint from the course below no longer reduces the hoop stress. For a locally thinned area, or any other point of interest, API 653 uses the full height measured from the bottom of the thinned area to the maximum liquid level, with no one-foot reduction.

Corroded areas are not judged on a single lowest reading. The code has the inspector find the least thickness in the area, average readings over a critical length derived from the tank diameter and that least thickness, and compare the average with t-min. The least thickness itself must also stay above a stated fraction of t-min. Widely scattered pits are handled separately: they can be ignored where no pit leaves less than half of t-min and the pits along any vertical line stay within a limited total length.

A worked example

Take a 100 ft diameter tank with a maximum liquid level 40 ft above the bottom of the first course, storing product with a specific gravity of 0.9. For illustration only, assume an allowable stress of 25,000 psi; the real value comes from Table 4.1 for the actual plate and course. Away from welds, with E = 1.0, the bottom course needs t-min = 2.6 x 39 x 100 x 0.9 / 25,000 = 0.365 in.

If the same location sits at a weld whose original joint efficiency is 0.85, t-min rises to 0.429 in. A locally thinned area whose lower edge is 2 ft above the course bottom has H = 38 ft and no one-foot reduction, giving 2.6 x 38 x 100 x 0.9 / 25,000 = 0.356 in., compared with the average thickness over its critical length. Changing to a product with a specific gravity of 1.0 raises every figure by about 11 percent, which is why the highest intended specific gravity, not the current product, belongs in the calculation.

Planning the ultrasonic survey so the calculation can be trusted

The calculation is only as good as the thickness data behind it. A Level III writing the survey procedure sets readings on every course, concentrated where corrosion is expected: the lowest course near the bottom, the liquid-vapour interface, around nozzles and wherever coating has failed. Spot readings find corrosion only where they happen to land, so suspect areas are scanned to find the least thickness and to map its extent for the averaging the code requires.

The procedure also controls the measurement itself: calibration on a reference block of known thickness in similar material, a suitable technique for coated plate, temperature correction where the shell runs hot, and records tied to course, orientation and height so that the next survey measures the same points. Readings that cannot be shown to come from the same locations cannot support a corrosion rate, and without a corrosion rate there is no defensible interval.

Common mistakes

Measuring H from the tank bottom for an upper course instead of from that course's bottom. Using nominal plate thickness instead of measured thickness. Taking the specific gravity of the current product when the tank may later store something heavier. Using the upper-course allowable stress on the bottom two courses. Setting E to 1.0 for a thinned area that sits on or next to a weld. Treating one low spot reading as the course thickness, or averaging away a deep local area without the least-thickness and pitting checks. Mixing units, particularly when a survey reports millimetres and the calculation is done in inches.

From t-min to the next inspection

The difference between measured thickness and t-min is the remaining corrosion allowance. Divided by the corrosion rate, it gives the remaining life, and API 653 sets the ultrasonic thickness interval for the shell from it: the lesser of half the remaining life and fifteen years where the rate is known, and five years where it is not. The API inspection interval calculator runs this arithmetic for screening; the authorized inspector, as inspector of record, makes the decision.

Atlantis NDT performs shell thickness surveys and supplies the readings to the owner's API 653 inspector; see API 653 tank inspection services. The API 653 summary covers the rest of the standard, and the tank interval guide explains how the inspection clocks interact.

Why does API 653 subtract one foot from the liquid height?

Because the whole-course check follows the one-foot design method, which evaluates each course at a point one foot above its bottom, where restraint from the course below no longer reduces the hoop stress. Locally thinned areas use the full height above them, without the reduction.

What is the minimum thickness of any API 653 shell course?

No course may be less than 0.1 in., whatever the calculation gives. On small or lightly loaded tanks the calculated t-min for the upper courses can fall below that figure, and the 0.1 in. floor then governs.

Which specific gravity goes into the API 653 shell calculation?

The highest specific gravity of the contents the tank is to store, not necessarily today's product. Moving a tank into heavier service raises t-min on every course and can turn an acceptable shell into one that needs repair or a lower fill height.

When can the joint efficiency be taken as 1.0 in API 653?

When the assessed area is far enough from any weld or joint, as defined in the shell evaluation section. At or near a weld, use the tank's original joint efficiency, or the code's table value when the original is unknown.

Can widely scattered pits be ignored in the shell evaluation?

Yes, within limits: no pit may leave less than half of t-min, and the pits along any vertical line must stay within the total length the code allows. Pitting outside those limits is evaluated as a corroded area.

Who signs off the shell thickness evaluation?

The authorized inspector certified to API 653, as inspector of record, with an engineer where the evaluation calls for one. NDT technicians take and record the readings under a written procedure; the acceptance decision belongs to the inspector.

Request a consultation

Need the NDE behind your API 653 programme?

Atlantis carries out tank floor, shell and settlement NDE and reviews the data; your Authorized Inspector stays inspector of record.

Next step: UT Level 1 course

The corrosion rates and remaining-life figures that API 653 works from come from ultrasonic thickness readings. The UT Level 1 course covers straight-beam thickness measurement, calibration on reference blocks and recording readings to a written procedure. Completing a course is not certification: under an employer-based scheme such as SNT-TC-1A, your employer certifies you after its own examinations.

For employers: ASNT Level III consulting covers procedures, the written practice, qualification examinations and audit support for your technicians.