Hydrotest Planning Under API 650 §7.3: Fill Rate, Settlement Monitoring, and the Shell Stress Check

The Hydrotest Basis: Fill Level and Test Medium

API 650 §7.3 requires that every new tank be hydrostatically tested with water — or another liquid that is compatible with the tank material and has a specific gravity no greater than 1.0 — filled to the design liquid level or to an agreed test level. The purpose of the hydrotest is to confirm the structural integrity of the shell, bottom, and roof welds under a load condition that is at least as severe as the maximum operating fill.

The test medium is almost always water (SG = 1.0). Using a lighter medium — for example, a partially salt-saturated water at SG = 1.02 for a seawater service tank — is possible but requires explicit agreement between the purchaser and the manufacturer. The test fill level must equal the design liquid level unless a lower level can be justified by the shell stress check (discussed in the next section). Underfilling to save time or reduce hydrostatic load on the foundation is not acceptable unless the hydrotest stress check specifically requires a reduced fill level.

For fixed-roof tanks with internal floating roofs, the hydrotest covers the shell and bottom only; the fixed roof is tested by a separate pneumatic test or visual examination of its welds. For open-top tanks and tanks without a fixed roof at the time of hydrotest, the shell and bottom test is the complete hydrostatic proof test.

Weld inspection must be complete before the hydrotest begins. All required visual examinations, radiographic tests, and vacuum box tests of the bottom plate lap welds must be accepted and recorded before water is introduced. The hydrotest is not a substitute for weld inspection — it is the final confirmation that an already-inspected tank holds water under design load.

The Hydrotest Shell Stress Check (The Often-Missed Calculation)

The hydrotest shell stress check is a separate calculation from the shell thickness design check, and it is the calculation most commonly omitted in practice. The two checks use different allowable stresses and different product specific gravities, and they can produce different governing courses.

For the design condition: each shell course is sized so that the hoop stress under the design product does not exceed the design allowable stress S_d. Per API 650 §5.6.1, S_d = 0.667 × SMYS (the specified minimum yield strength of the shell material). The product SG specified in the design data is used directly in the thickness formula.

For the hydrotest condition: the tank is filled with water (SG = 1.0) to the full test level, regardless of the actual product SG. The allowable stress for the hydrotest is S_t = 0.75 × SMYS — 12.5% higher than S_d. This higher allowable is justified because the hydrotest is a short-duration proof load, not a continuous service condition.

The consequence: if a tank is designed for a light product (say SG = 0.75 — a condensate service), the shell courses were sized for 75% of the hydrostatic head that water would produce. When the tank is flooded with water for the hydrotest, the actual hoop stress in the lower shell courses is 1.0/0.75 = 33% higher than the design stress. Whether this exceeds S_t = 0.75 × SMYS must be checked explicitly for every course.

If any shell course has a hydrotest hoop stress exceeding S_t, the test fill level must be reduced to a level at which all courses remain within S_t. This is a partial fill hydrotest. The maximum allowed test fill height H_t for the limiting course is found by rearranging the hoop stress formula to solve for H at the point where σ = S_t.

If a partial fill hydrotest is required, nozzles and attachments above that reduced water level may need separate code-compliant testing or verification under the approved test procedure. The shell fill height alone does not automatically complete every hydrotest obligation on the tank.

Partial fill hydrotests are common for tanks storing light products (SG below about 0.80) and for tanks built from lower-grade steels where SMYS is relatively low. They are not a deficiency — they are a planned outcome of the design process. What causes project problems is discovering the partial fill requirement on the day of the test rather than weeks in advance when the test procedure is being written.

Fill Rate and Settlement Monitoring Intervals

API 650 §7.3.4 requires controlled filling — the tank shall not be filled at a rate that prevents meaningful settlement monitoring. The standard does not state an explicit maximum fill rate in millimetres per hour, but common practice, consistent with API 651 guidance, is to limit the fill rate to 300 mm of water height per hour during the first fill of a new tank.

The engineering rationale is straightforward: granular foundation materials (crushed stone, compacted sand) redistribute stress as load is applied incrementally. If filling is too rapid, the foundation load increases faster than the material can reposition, and localised differential settlement can develop and stabilise only after the tank has been filled past the point where it could be caught and evaluated. Slow filling allows the foundation to consolidate under each load increment before the next is applied.

Beyond settlement physics, a controlled fill rate gives the inspection team time to take and record settlement readings at every specified hold point. Rushing the fill to meet a commissioning deadline — and taking compressed or estimated readings — produces a settlement record that is useless for future API 653 assessments. The initial hydrotest settlement survey is the baseline against which all future in-service surveys are compared.

Mandatory hold points for settlement readings must be specified in the test procedure before filling begins:

• Baseline (pre-fill): All perimeter settlement markers and shell plumb readings recorded with zero water in the tank.
• Every 3 m of water height (or every 25% of the tank shell height, whichever is more frequent): hold filling, allow a minimum stabilisation period (typically 1–2 hours for granular foundations, longer for soft clay), then take full settlement readings before resuming.
• At full test fill level: A full survey with a minimum 24-hour hold at full fill before the final settlement survey is taken.
• After dewatering: Final post-test survey to assess rebound and permanent set.

Allowable Settlement: What API 650 Appendix B Says

API 650 Appendix B provides the framework for evaluating foundation settlement. Three distinct modes of settlement are recognised, each with a different allowable limit and a different structural consequence.

Uniform settlement — the entire tank sinks as a rigid body without any differential movement between points. This is structurally benign: the tank, its nozzles, and its piping all move together. API 650 Appendix B places no limit on uniform settlement from a tank structural perspective (piping flexibility may govern instead). In practice, uniform settlements of 50–150 mm are common on soft ground and are acceptable provided piping is flexible enough to follow the tank.

Planar tilt (differential settlement across the tank diameter) — one side of the tank sinks more than the other, causing the tank to tilt as a whole. The structural concern is the eccentric hydrostatic loading on the shell and the potential overstress of the shell-to-bottom weld on the high side. API 650 Appendix B limits the tank tilt angle θ to:

θ ≤ arctan(50 / D)

where D is the tank diameter in metres. For a 30 m diameter tank this gives a maximum tilt angle of approximately 1.67° (equivalent to about 875 mm difference in settlement between opposite sides of the shell). For a 60 m diameter tank the limit is approximately 0.84°. Note that the arctan(50/D) limit is measured from the perimeter markers — the difference between the highest and lowest perimeter readings divided by the diameter gives tan(θ).

Local or point settlement — a single shell plate or small arc of the shell foundation sinks more than adjacent areas, creating a local low point in the bottom. This is the most structurally damaging settlement mode because it concentrates bending at the shell-to-bottom weld over a short arc length. API 650 Appendix B limits local differential settlement to 13 mm over any 3 m arc of the perimeter. Where markers are spaced at every third shell plate (approximately 6 m spacing for typical plate widths), interpolation or intermediate readings may be needed to confirm compliance with the 3 m arc limit.

If settlement readings taken during filling approach either the tilt limit or the local settlement limit, filling must be stopped and the foundation assessed before proceeding. Continuing to fill a tank that is already at the settlement limit can cause permanent bottom distortion or shell-to-bottom weld cracking — both of which require expensive remediation.

What to Document Before, During, and After

The hydrotest documentation package is a permanent record that travels with the tank through its operational life. It is reviewed at every API 653 inspection and is referenced whenever a change of service, a nozzle modification, or a re-rating is considered. A complete package eliminates ambiguity about whether settlement readings were taken, what the foundation condition was at initial fill, and whether the test met the standard.

Before the hydrotest:

• Weld inspection records — visual examination, radiographic test results, magnetic particle or dye penetrant results, and vacuum box test records for all bottom plate lap welds. All outstanding punch items must be cleared and signed off.
• Calibration records for all measurement instruments: survey level, theodolite or total station (for shell plumb), thermometer (water temperature for density correction if using non-standard SG), and elevation measurement points.
• The written hydrotest procedure: approved fill rate, hold point elevations, settlement marker locations and labelling convention, allowable settlement limits (from the design calculation), responsible inspector identity, and the acceptance criteria for the test.
• The hydrotest shell stress check calculation for every course, signed and dated, showing the maximum test fill level. If a partial fill is required, the test procedure must state the maximum fill height explicitly.
• Baseline settlement survey: elevation of all perimeter markers and shell plumb readings (vertical deviation at top of each shell plate) recorded at zero fill.

During the hydrotest:

• Fill log: a dated and timestamped record of water height at regular intervals (every hour minimum), showing the fill rate at each interval. Any deviation from the planned fill rate must be noted with the reason.
• Settlement readings at every hold point: elevation of each perimeter marker, calculated differential settlement from baseline, and a notation confirming whether the readings are within allowable limits before filling resumed.
• Any anomalies observed: weld seepage (note location, time, and whether seepage stopped or continued), bottom plate movement, unusual sounds.

After the hydrotest:

• Final settlement survey at full fill: complete perimeter elevation survey and shell plumb readings, with pass/fail notation against the allowable limits.
• Post-dewatering survey: perimeter elevations and shell plumb readings after the tank has been fully drained, to record permanent set and rebound.
• Sign-off sheet: inspector name and company, date, statement of pass or conditional pass with any noted observations, and the responsible engineer's certification that the test was completed in accordance with API 650 §7.3 and the approved test procedure.

Practical Tips for Safe Hydrotesting

1. Perform the hydrotest shell stress check before scheduling the hydrotest date. If a partial fill is required, this must be communicated to the client, the civil engineer (who needs to assess the reduced foundation load and confirm whether it still provides adequate testing of the soil), and the NDE team (who need to know the fill level to plan access for final weld inspections during fill). Discovering a partial fill requirement the day before the hydrotest creates contractual disputes and schedule delays that are entirely avoidable.
2. Label perimeter settlement markers before the baseline survey. Use a consistent convention — for example, clockwise from north, every third shell plate — and photograph the markers before filling. Ambiguous marker labels are one of the most common causes of unusable settlement records.
3. For tanks on ring-wall foundations, confirm whether the ring-wall geotechnical report specifies a maximum fill rate. The geotechnical engineer may have imposed a more restrictive fill rate than the API 650 default 300 mm/hour based on the bearing capacity of the subgrade. If a geotechnical report exists, read it before writing the hydrotest procedure.
4. Plan the dewatering route in advance. A 30 m diameter tank filled to 15 m holds approximately 10,600 m³ of water. Dewatering this volume must be directed to an acceptable discharge point — storm drains, onsite retention, or a licensed disposal route. Dewatering too quickly (by gravity through a single 150 mm drain nozzle) can also cause bottom plate distortion from suction if the roof is sealed — ensure adequate ventilation during dewatering.
5. If the tank has a floating roof, the hydrotest also tests the floating roof pontoons and seal for water-tightness. The floating roof must be floating freely at the full fill level; any pontoon compartments showing ingress of water must be repaired before the tank is placed in service. Record the floating roof position and condition at full fill as part of the hydrotest record.
6. In cold climates, hydrotest water temperature matters. Do not hydrotest when ambient temperatures are forecast to drop below 4°C during the test hold period. Freezing water in the tank can cause catastrophic shell failure and permanent bottom distortion. If cold-weather hydrotesting is unavoidable, heat the fill water and monitor its temperature continuously.

Calculate the hydrotest shell stress before the hydrotest date, not the day before. A partial fill hydrotest requires coordination with the client, the civil engineer (reduced load on foundation), and the NDE team. Discovering it late derails the construction schedule.

Related reading: Continue with Shell Course Thickness Design, Foundation Loads for API 650 Tanks, and Corrosion Allowance Strategy to keep the full API 650 design workflow connected.