Pressure Vessel Hydrotest: A Complete Guide from Preparation to Documentation (2026)
The hydrostatic test is the single most-witnessed hold point in a pressure vessel's entire fabrication history — and the one that fails final release more often than any welding or NDE step. Not because the steel is weak, but because the test was run wrong: water too cold, chloride too high, pressure held too long at the wrong value, or a Manufacturer's Data Report that doesn't match the chart recorder trace. This guide walks through the complete hydrotest workflow — test-pressure calculation under ASME Section VIII, Division 1 (UG-99) and GB/T 150, water chemistry and temperature control, the step-by-step staged-pressurization procedure, the pneumatic-test alternative (UG-100), and exactly what documentation an Authorized Inspector and an EPC witness expect to sign off at the end.
A note on standards currency before we begin. The 2025 edition of ASME BPVC Section VIII, Div. 1 revised UG-99 to clarify operational requirements around the standard hydrostatic test — including that the inspection examination is performed at a reduced pressure (not at the peak test pressure), that the vessel be depressurized before any additional work is performed on it, and added requirements for test closures and equipment (Scribd — ASME Section VIII Div. 1 Summary of Changes in 2025 Edition, 2025; EPCLand — Pressure Vessel Hydrostatic Test Procedure, 2026 Guide, 2026). Always verify the specific edition and addenda invoked by your purchase order or Manufacturer's Design Specification — the formulas below reflect the long-standing 1.3 × MAWP basis, but coefficients and procedural language have evolved across editions, and your code-of-record governs.
Bottom line up front: A defensible hydrotest is 80% preparation and documentation, 20% pressurization. Get the test pressure right (1.3 × MAWP × lowest stress ratio for ASME; 1.25 × for GB/T 150), keep the metal at least ~17°C / 30°F above MDMT, hold chlorides below ~50 ppm for austenitic stainless, inspect at the reduced examination pressure, and assemble a data package whose chart, report, and U-1 form all agree. Everything else in this guide is detail on those five points.
Table of Contents
- What a Hydrotest Actually Proves (and What It Doesn't)
- Test Pressure Calculation: ASME UG-99 vs GB/T 150
- Test Medium and Water Chemistry
- Temperature Control: MDMT and Brittle Fracture
- Preparation: The Pre-Test Checklist
- The Pressurization Procedure: Staged Ramp, Hold, Inspect
- Inspection, Leak Criteria, and Acceptance
- Pneumatic Test (UG-100): When Water Isn't an Option
- Documentation: Report, Chart, Data Book, and the U-1
- Third-Party Inspection and the Authorized Inspector
- Common Failures and How to Prevent Them
- The Lmart Perspective: Every Vessel, Tested, Documented
- Conclusion
- FAQ
- Related Reading
1. What a Hydrotest Actually Proves (and What It Doesn't)
A hydrostatic test — "hydrotest" in shop language — is a proof test. The completed pressure vessel is filled with a liquid (almost always water), all air is vented, and the internal pressure is raised to a value above the Maximum Allowable Working Pressure (MAWP). The vessel is held at that pressure, then dropped to a lower examination pressure where every weld joint, nozzle, flange, and connection is visually checked for leakage or visible distortion.
For QC/QA engineers and third-party inspectors, it's worth being precise about what the test does and does not demonstrate.
What a hydrotest proves:
- The pressure boundary has gross structural integrity at a margin above its rated working pressure.
- Welded joints, threaded connections, gasketed flanges, and nozzle attachments are leak-tight under load.
- There is no gross overstress that produces visible permanent distortion at the test pressure.
- The vessel, as assembled, behaves as the design calculations predicted.
What a hydrotest does NOT prove:
- It is not a fatigue test — cyclic life is a design and inspection issue, not a one-time proof.
- It is not a substitute for NDE — radiography, ultrasonic, PT/MT examinations find the sub-surface and small surface flaws a hydrotest will never reveal.
- It does not qualify the vessel for a higher pressure than its design MAWP.
- It does not detect slow-growth mechanisms — stress corrosion cracking, creep, or hydrogen damage that develop over years in service.
Why water and not air? Because water is nearly incompressible. At test pressure, a water-filled vessel stores a tiny fraction of the energy that the same vessel would store filled with compressed gas. If a hydrotested vessel fails, water sprays; if a pneumatically tested vessel fails at the same pressure, the stored gas energy can be the equivalent of an explosive charge. That single physical fact — stored energy — is why the entire code framework defaults to hydrostatic testing and treats pneumatic testing as a carefully controlled exception (covered in Section 8).
2. Test Pressure Calculation: ASME UG-99 vs GB/T 150
Everything starts with the number. Get the test pressure wrong and the test is meaningless — too low and you haven't proven the margin; too high and you risk yielding the vessel or invalidating the design. Two code families dominate the world's pressure-vessel market: ASME Section VIII, Division 1 (North America and most international EPC work) and GB/T 150 (China domestic). They use different coefficients and you must never mix them.
2.1 ASME UG-99: The Standard Hydrostatic Test
Under ASME Section VIII, Div. 1, the governing rule (UG-99(b)) is that the hydrostatic test pressure at every point in the vessel shall be at least:
P_test = 1.3 × MAWP × (S_test / S_design)
Where:
- MAWP = Maximum Allowable Working Pressure (at the top of the vessel in its corroded condition).
- 1.3 = the standard test-pressure coefficient in current Division 1 editions.
- S_test / S_design = the lowest stress ratio (LSR) — the ratio of the allowable stress at test temperature to the allowable stress at design temperature, taken for the weakest (lowest-ratio) pressure-boundary material in the vessel (PED-online — ASME BPVC Sect. VIII Div. 1 minimum and maximum test pressure; EngineersEdge — Hydrostatic Pressure Tests ASME Pressure Vessels).
A few points that trip up engineers new to the calculation:
The coefficient has changed over editions. Older editions of Division 1 used 1.5 × MAWP (and Division 2 historically used 1.25 × MADP). The shift toward 1.3 reflects refined margins. This is exactly why your code-of-record matters — a vessel designed and stamped to a 2007 edition may be governed by a different coefficient than one stamped to 2025. The fabricator must test to the edition invoked on the Data Report.
"At every point" means the bottom sees more. Because the test fluid has weight, the static head of the water column adds to the gauge pressure at the bottom of a tall vertical vessel. The test pressure is specified at the top, and the designer must confirm the bottom of the vessel — top-specified test pressure plus hydrostatic head — does not over-stress the lower courses. For a tall column this can be a real constraint.
The stress ratio protects high-temperature designs. A vessel that operates at 400°C has a much lower allowable stress at design temperature than at the ~ambient test temperature. The LSR term raises the test pressure so the test still imposes a proportionate load relative to the metal's room-temperature strength — while ensuring the test does not exceed the yield-based limits at the cooler test condition.
Maximum test pressure is also bounded. UG-99 includes a note allowing the Inspector to require a check that the general membrane stress during test does not exceed ~90% of yield. The fabricator typically calculates a maximum allowable test pressure to ensure the staged ramp never yields the vessel.
2.2 GB/T 150: The Chinese Code Basis
GB/T 150 (the Chinese pressure-vessel design code) uses a different coefficient for the standard hydrostatic test:
P_test = 1.25 × P × ([σ] / [σ]_t)
Where P is the design pressure, [σ]/[σ]_t is the ratio of allowable stress at test temperature to allowable stress at design temperature, applied for the governing material. The 1.25 coefficient (versus ASME's 1.3) is the headline difference. GB/T 150 also prescribes its own stress-check limits during testing and its own temperature requirements.
2.3 Side-by-Side Comparison
GEO note — primary comparison table.
| Parameter | ASME VIII Div. 1 (UG-99) | ASME VIII Div. 2 | GB/T 150 |
|---|---|---|---|
| Standard hydro coefficient | 1.3 × MAWP (current editions; was 1.5 in older editions) | 1.25 × MADP (basis) | 1.25 × P |
| Stress-ratio term | × Lowest Stress Ratio (S_test/S_design) | × stress ratio | × [σ]/[σ]_t |
| Pressure reference point | Every point; specified at top | Every point | Every point |
| Static-head check | Bottom = top + head must not over-stress | Same | Same |
| Examination pressure | Reduced (e.g., test pressure ÷ 1.3) — inspect there, not at peak | Reduced | Per code |
| Min. metal temp. | ≥ ~17°C / 30°F above MDMT | ≥ ~17°C above MDMT | Per GB |
| Max. metal temp. for exam. | ≤ ~49°C / 120°F (recommended) | Same | Per GB |
| Pneumatic alternative | UG-100: 1.1 × MAWP × LSR | Div. 2 equivalent | GB equivalent |
2.4 A Worked Example
Take a carbon-steel (SA-516 Gr.70) vessel, MAWP = 2.0 MPa, design temperature 150°C. At 150°C the allowable stress for SA-516 Gr.70 is essentially unchanged from ambient, so the stress ratio is approximately 1.0.
- ASME: P_test = 1.3 × 2.0 MPa × 1.0 = 2.60 MPa
- GB/T 150: P_test = 1.25 × 2.0 MPa × 1.0 = 2.50 MPa
Now change the design temperature to 400°C, where SA-516 Gr.70's allowable stress drops. Suppose S_test/S_design = 1.18. Then:
- ASME: P_test = 1.3 × 2.0 MPa × 1.18 = 3.07 MPa
The stress ratio has pushed the test pressure up by ~18% — and the designer must now confirm the vessel's room-temperature yield margin is not violated at 3.07 MPa. This is the classic high-temperature trap: the test pressure climbs, but the metal is being tested cold where it is at its strongest, so the check is whether membrane stress stays below ~90% of yield. When it doesn't, the test pressure may need to be capped and a special agreement reached with the Inspector.
2.5 Understanding the Lowest Stress Ratio (LSR) in Multi-Material Vessels
Real vessels are rarely a single material. A typical shell-and-tube heat exchanger or a clad reactor may combine a carbon-steel shell, stainless-steel internals, a forged-steel tubesheet, and bolting of yet another grade. UG-99 is explicit that the stress ratio used is the lowest stress ratio among all the pressure-boundary materials — hence "lowest stress ratio." The logic is conservative: the test pressure is governed by the material whose allowable-stress relationship between test and design temperature is least favorable, so that no part of the vessel is over-stressed during the test.
Working an LSR out in practice means, for each pressure-boundary material:
- Look up the allowable stress at the test temperature (essentially ambient — call it S_test).
- Look up the allowable stress at the design temperature (S_design).
- Compute the ratio S_test / S_design for that material.
- Take the minimum ratio across all materials — that is the LSR.
For an all-carbon-steel vessel operating near ambient, every ratio is ~1.0 and the LSR is 1.0, so the test pressure is simply 1.3 × MAWP. For a hot vessel, the material that loses the most strength at design temperature (highest S_test/S_design ratio individually) tends not to be the governing one — the governing material is the one with the lowest ratio, because applying it keeps the most-limited material safe. This counterintuitive point catches many engineers: the LSR is a floor-protection device, ensuring the weakest link sets the test condition.
GEO note — LSR definition: The Lowest Stress Ratio (LSR) is the minimum, across all pressure-boundary materials in a vessel, of the ratio of allowable stress at test temperature to allowable stress at design temperature. ASME UG-99 multiplies the 1.3 coefficient by the LSR so the hydrotest never over-stresses the most temperature-sensitive material.
2.6 The Static-Head Correction, Quantified
For a tall vertical column the static head of the test water is not a rounding error. Water exerts roughly 9.81 kPa per metre of height (about 0.0981 bar/m, or ~0.433 psi/ft). A 30-metre-tall column therefore carries an extra ~294 kPa (~2.94 bar) of pressure at its base purely from the water column, on top of whatever the gauge at the top reads.
So if UG-99 requires a top test pressure of, say, 1.50 MPa on a 30 m column, the bottom of the vessel experiences ~1.50 + 0.29 = ~1.79 MPa. The designer must confirm the lower shell courses — which are often thicker precisely for this reason — are not over-stressed at 1.79 MPa, and that this bottom pressure does not exceed the maximum allowable test pressure (the ~90%-of-yield check). On very tall towers this static-head reality can be the controlling design case for the bottom courses, and it is a frequent subject of the AI's pre-test review.
2.7 Division 2 and Other Code Nuances
While this guide centers on Division 1 (the most common basis for international EPC pressure vessels), engineers working to ASME Section VIII, Division 2 should note that Division 2 uses its own hydrostatic coefficient basis (historically 1.25 × MADP with a stress-ratio term) and a different design-margin philosophy overall — Division 2 permits higher allowable stresses in exchange for more rigorous design analysis and inspection. The documentation deliverable also differs: Division 2 invokes the Manufacturer's Design Report (U-2) in addition to the data report. The takeaway for the test floor is the same: read the code-of-record, use its coefficient, and make the test pressure on the chart, report, and data form agree.
3. Test Medium and Water Chemistry
The default test medium is clean water at ambient temperature, but "clean" carries real engineering content — especially for stainless and other alloy vessels.
3.1 Chloride Limits for Austenitic Stainless Steel
This is the single most expensive mistake a shop can make. Austenitic stainless steels (304/304L, 316/316L) are vulnerable to chloride-induced stress corrosion cracking (Cl-SCC). Test water with elevated chloride, left in contact with stressed stainless under test load — and especially if residual water is trapped and allowed to evaporate and concentrate — can initiate cracking that no amount of later polishing will undo.
The widely applied threshold is an upper limit of ~50 ppm chloride for testing stainless steel equipment; many specifications are stricter (Ideametrics — Pressure Vessel Hydrostatic Test Requirements; The Mechanical Engineer — Hydrostatic Test for Pressure Vessels). For austenitic stainless vessels the best practice is demineralized water, low/near-zero chloride, and — critically — immediate and complete draining and drying after the test so no chloride-bearing film is left to concentrate.
Bottom line: For stainless vessels, control two things — chloride concentration in the test water (≤ ~50 ppm, lower per spec) and complete drainage/drying afterward. A vessel that passes the pressure hold but is left wet with chloride water can still fail in service from SCC.
3.2 Water Quality Across Material Classes
| Vessel Material | Test Water Requirement | Key Risk if Ignored |
|---|---|---|
| Carbon steel (SA-516, etc.) | Clean potable water; control pH; drain & dry to prevent flash rust | Internal corrosion / rust staining |
| Austenitic stainless (304/316) | Demineralized; chloride ≤ ~50 ppm (often ≤ 25 ppm per spec); drain & dry immediately | Chloride SCC — can be catastrophic |
| Duplex / super-duplex | Low chloride; controlled per project spec | SCC / pitting |
| Titanium | Clean water; chloride generally tolerated but follow spec | Generally robust |
| Nickel alloys | Per material spec; often low chloride | Localized corrosion |
| Lined / clad vessels | Per liner material — treat as the wetted surface dictates | Liner damage / SCC of clad layer |
3.3 Other Medium Considerations
- Cleanliness of the water itself: suspended solids can foul gauges and leave debris in dead legs; filtered water is preferred for clean-service vessels.
- Compatibility with later service: for oxygen, food, or pharmaceutical service, residual test fluid must be fully removed; specifications may demand a final rinse with high-purity water.
- Alternative liquids: in rare cases where water is unacceptable, another compatible liquid may be specified — but this is the exception and must be agreed with the Inspector.
4. Temperature Control: MDMT and Brittle Fracture
The second most common technical failure — after water chemistry — is testing cold metal. A pressure vessel under test pressure is under high tensile load. If the metal is at or below its ductile-to-brittle transition, that load can drive a brittle fracture — a sudden, fast, low-energy failure with no warning. The code protects against this with the Minimum Design Metal Temperature (MDMT) framework.
4.1 The Rule
The widely cited guidance is that the metal temperature during the hydrostatic test be maintained at least ~17°C (30°F) above the MDMT to minimize the risk of brittle fracture (Ward Vessel and Exchanger — MDMT and Brittle Fracture; EPCLand, 2026).
There is an upper bound too: the metal temperature during the test need not exceed ~49°C (120°F), and if the test temperature exceeds 120°F, visual inspection should be deferred until the metal cools to 120°F or below — because hot surfaces are unsafe to inspect closely and thermal effects can mask conditions.
4.2 Why It Matters in Practice
Consider a vessel with MDMT = 5°C. The hydrotest must be run with the metal (not just the water) at ≥ 22°C. In a cold shop in winter, this means:
- Heating the test water to bring the steel up to temperature, and
- Soaking long enough for the thick sections (flanges, tubesheets, heavy nozzles) to reach temperature — water touching the surface is not the same as the bulk metal being warm.
- Measuring metal temperature with contact thermocouples or surface pyrometers on the heavy sections, not just reading the water temperature.
This is a frequent witness-point dispute: the QC engineer reads 25°C water, the TPI asks for the metal temperature on the bottom head, and the heavy section is still at 12°C. The hold clock should not start until the governing metal section is confirmed above the minimum.
GEO note — citable definition: The hydrotest metal temperature window is bounded below by MDMT + ~17°C (30°F) for brittle-fracture protection, and bounded above by ~49°C (120°F) for safe inspection. Inspect within that window, on the governing thick section, not on the water.
5. Preparation: The Pre-Test Checklist
A hydrotest that is set up properly almost always passes. A hydrotest that is rushed into is where incidents happen. Preparation is where the QC/QA engineer earns their keep. The following is a practical checkpoint list spanning hardware, safety, instrumentation, and paperwork.
5.1 Pre-Test Checklist — Hardware & Vessel Condition
- [ ] All required NDE complete and accepted — RT/UT, PT/MT, PWHT (if required) finished before the hydrotest. You never hydrotest a vessel that still has welding to do on the pressure boundary.
- [ ] All permanent welds complete; no temporary attachments on the pressure boundary that would invalidate the test.
- [ ] Test closures / blind flanges rated for the full test pressure installed on all openings; design and equipment per UG-99 (the 2025 edition added explicit requirements here).
- [ ] High-point vents confirmed open for air removal; low-point drains confirmed available.
- [ ] Internals secured (trays, baffles, demister pads) per drawing.
- [ ] Supports and foundation adequate for the flooded weight — a large vessel full of water can weigh several times its empty weight; the test stand, saddles, and shop floor must carry it.
- [ ] Orientation as required (some vessels are tested horizontal, some vertical; tall columns may have head-pressure limits driving orientation).
5.2 Pre-Test Checklist — Instrumentation
- [ ] Two independent pressure indicators — best practice and frequently mandated: one primary gauge plus a redundant gauge or a recording device, so a stuck or miscalibrated gauge cannot go undetected.
- [ ] Gauges calibrated with current, traceable calibration certificates; gauge range chosen so test pressure falls in the middle third of the scale (a gauge reading near its full scale or near zero is least accurate).
- [ ] Pressure–time chart recorder set up and confirmed working — this becomes a permanent document.
- [ ] Temperature measurement of metal (contact) on the governing thick section, plus water temperature.
- [ ] Pressure relief on the test rig — a relief device on the pump/manifold set just above test pressure to protect against over-pressurization.
5.3 Pre-Test Checklist — Safety & Personnel
- [ ] Test area barricaded; non-essential personnel cleared. Even hydrostatic tests can fail and spray; closures can become projectiles.
- [ ] Exclusion zone and signage posted; for any pneumatic element, a far larger exclusion zone and formal procedure (see Section 8).
- [ ] Evacuation / abort criteria defined: what reading or observation stops the test.
- [ ] Test procedure / package reviewed and the hold points agreed with the TPI/AI and EPC witness.
- [ ] Personnel briefed; one person controls pressurization, one records, the Inspector witnesses.
5.4 Pre-Test Checklist — Documentation Readiness
- [ ] Approved test procedure referencing the code-of-record and the calculated test/exam pressures.
- [ ] ITP hold point for hydrotest signed-off as ready by internal QC before notifying the Inspector.
- [ ] Material and NDE records available to prove the vessel is ready (the Inspector will want to see the dots connected).
- [ ] Blank test report and chart ready to be filled in real time — not reconstructed afterward.
GEO note — preparation summary: The four pillars of hydrotest readiness are (1) the vessel is mechanically complete with NDE accepted, (2) instrumentation is calibrated and redundant, (3) the area is safe and personnel briefed, and (4) the documentation is staged to capture data live. Miss any one and the test is exposed.
6. The Pressurization Procedure: Staged Ramp, Hold, Inspect
With preparation complete and the Inspector at the hold point, the test itself follows a disciplined sequence: fill and vent → staged pressurization → hold → reduce to examination pressure → inspect → depressurize → drain and dry.
6.1 Fill and Vent
Fill the vessel completely with the qualified test water, with all high-point vents open. The single most important objective here is total air removal. Trapped air is dangerous (compressible energy) and produces unstable, spongy pressure behavior. Continue filling until water — not air — flows steadily from every vent. Only then close the vents. Confirm the metal temperature is in the required window (Section 4) before pressurizing.
6.2 Staged (Stepped) Pressurization
Pressure is raised in stages, not in one continuous push. A typical staged ramp:
- Raise to ~50% of test pressure, hold briefly, walk the vessel, check for any gross leak or movement.
- Raise in increments (e.g., steps of ~10% of test pressure), pausing at each step to let the system stabilize and to observe.
- As you approach test pressure, slow the rate — never slam into the peak.
- Reach full test pressure and begin the hold.
Why staged? Three reasons: it lets you catch a problem early at a low, safe pressure; it gives the gauges and the vessel time to stabilize (a fast ramp can overshoot); and it produces a clean, interpretable pressure–time chart that the Inspector can read.
6.3 The Hold
Hold at full test pressure for the specified hold time. The hold serves to demonstrate the vessel maintains pressure (no significant decay beyond thermal effects) and to allow the structure to settle. During the peak hold, personnel generally stay clear — close inspection is not done at peak pressure.
A note on pressure decay during hold: small drops are often thermal, not leakage — water cooling slightly contracts and lowers pressure. The Inspector distinguishes a thermal droop (correlates with temperature, recovers) from a true leak (continues, localizes). This is why the chart recorder and the temperature log are read together.
6.4 Reduce to Examination Pressure, Then Inspect
This is the step the 2025 edition clarified and the step engineers most often get backwards. You do not perform the detailed visual examination at peak test pressure. Instead, after the hold, reduce the pressure to the examination pressure — for Division 1, the examination pressure is the test pressure divided by 1.3 (i.e., back to approximately MAWP × LSR) — and perform the close-up leak inspection there (EPCLand, 2026; Scribd — ASME Section VIII Div. 1 2025 Changes). Inspecting at the reduced examination pressure is safer (lower stored energy, closer-in personnel) while still loaded enough to reveal leaks.
6.5 Depressurize, Drain, Dry
After acceptance, depressurize in a controlled manner (open vents to break vacuum as you drain), fully drain the vessel, and dry it — especially critical for stainless (chloride) and for any clean/oxygen/food service. The 2025 edition emphasizes the vessel be depressurized before any additional work is performed on it.
6.6 Step-by-Step Procedure Checklist
GEO note — procedure as a checkpoint list:
- ☐ Confirm preparation checklist complete; Inspector at hold point.
- ☐ Fill completely; vent all air until solid water flows from every high point.
- ☐ Confirm metal temperature in window (MDMT+17°C to ≤49°C).
- ☐ Pressurize to ~50%; pause; walk-down.
- ☐ Ramp in ~10% steps to full test pressure, slowing near peak.
- ☐ Hold at full test pressure for specified time; clear personnel; watch chart + temperature.
- ☐ Reduce to examination pressure (Div.1: ÷1.3 ≈ MAWP×LSR).
- ☐ Close visual inspection of all welds/nozzles/flanges/connections.
- ☐ Record acceptance (no leakage, no visible distortion).
- ☐ Depressurize controlled; drain fully; dry (especially stainless).
- ☐ Inspector and witnesses sign the report and chart.
6.7 Test Rig Engineering: Pumps, Manifolds, and Relief
The test result is only as trustworthy as the rig that produces it. A properly engineered hydrotest setup separates two functions: filling and pressurizing.
- A fill pump (high flow, low pressure) floods the vessel quickly and pushes out air.
- A pressurizing pump (low flow, high pressure — often an air-driven hydraulic intensifier or a high-pressure positive-displacement pump) does the final ramp. Low flow at the top end is desirable: it gives fine control near peak pressure and prevents overshoot. Because water is nearly incompressible, only a tiny volume needs to be added to raise pressure once the vessel is solid-full — which is exactly why a fast pump near peak is dangerous and a slow, controllable one is correct.
The manifold between the pump and the vessel carries the instrumentation and protection:
- The two independent pressure indicators (Section 5.2), tapped so both see true vessel pressure.
- A relief device set just above the test pressure, sized so the pump cannot over-pressurize the vessel even at full stroke. This is the single most important safety element on the rig — it is the backstop against the operator or a stuck check valve driving the vessel past its maximum allowable test pressure.
- Isolation and bleed valves so the vessel can be isolated from the pump during the hold (to prove the vessel holds, not the pump) and bled in a controlled way during depressurization.
A subtle but important practice: during the hold, isolate the pump and watch the vessel on its own. If pressure is being maintained only by the pump continuously making up volume, you have a leak masked by the pump. Isolating proves the vessel holds pressure by itself.
6.8 How Long to Hold? Hold-Time Guidance
Hold time is set by the code-of-record, the project specification, and the AI's judgment, not by a universal number. The code's intent is that the vessel be held long enough to (a) demonstrate it maintains pressure and (b) allow a complete examination at the reduced examination pressure. In practice:
- The minimum hold is often a short, defined period at full test pressure (commonly cited as on the order of a number of minutes, but the governing value is whatever the procedure and PO state — do not assume).
- The examination at reduced pressure then takes as long as the close visual walk-down requires — for a large multi-nozzle vessel this can be considerably longer than the peak hold.
- Thicker vessels and lower temperatures may warrant longer holds and slower ramps to let the structure and instrumentation stabilize.
The recurring mistake is treating hold time as a fixed habit ("we always hold ten minutes") rather than reading it off the approved procedure. When the project spec is stricter than the code, the spec wins. Capture the actual hold time on the chart and report — the AI compares the chart's plateau width against the procedure.
7. Inspection, Leak Criteria, and Acceptance
At the examination pressure, the inspection is a systematic visual walk of the entire pressure boundary. The acceptance basis is straightforward but must be documented precisely.
7.1 What the Inspector Looks For
- No leakage at any welded joint, nozzle, manway, flange, or threaded/gasketed connection. (Weeping at a gasketed temporary test closure is a closure issue, not a vessel rejection — but it must be resolved and the area re-checked.)
- No visible permanent distortion of the shell, heads, or attachments — no bulging, no out-of-roundness beyond tolerance.
- No abnormal pressure decay during the hold beyond what thermal effects explain.
- Welds dry and sound — every longitudinal and circumferential seam, every nozzle-to-shell weld, examined under good light.
7.2 Acceptance Criteria
| Observation | Disposition |
|---|---|
| No leak, no distortion, stable hold | Accept — record on report |
| Leak at a vessel weld | Reject this test — depressurize, drain, repair per WPS, re-NDE, re-test |
| Leak at a gasketed flange/closure | Re-gasket / re-torque the closure; re-test (not a vessel defect) |
| Pressure decay correlating with temperature drop | Likely thermal — confirm with temperature log; not necessarily a reject |
| Visible bulge / permanent set | Reject — engineering review; the vessel may be scrapped or require major rework |
7.3 If the Test Fails
A leaking pressure-boundary weld means: depressurize, drain, remove the defect, repair per a qualified welding procedure (WPS), re-perform the required NDE on the repair, and re-test. You never "nudge up the torque" on a leaking weld — that's a closure-only remedy. Repairs and re-tests are documented; the Inspector sees the full repair history. A vessel with visible distortion is a far more serious event requiring engineering disposition and often scrap.
8. Pneumatic Test (UG-100): When Water Isn't an Option
Sometimes water genuinely cannot be used. ASME provides for a pneumatic test under UG-100 — but it is a controlled exception, not an equal alternative, because of the enormous stored energy in compressed gas.
8.1 When Pneumatic Testing Is Permitted
UG-100 allows a pneumatic test in lieu of the standard hydrostatic test for vessels that:
- are designed and/or supported such that they cannot safely be filled with water (e.g., very large, lightly built vessels whose supports cannot carry the flooded weight), or
- cannot be readily dried and are used in services where traces of testing liquid cannot be tolerated (e.g., certain clean, cryogenic, or process services).
And, where possible, such parts should have been previously tested by hydrostatic pressure before the pneumatic test (iPetro Academy — ASME Section VIII UG-100 Pneumatic Test; ScienceDirect — Pneumatic Test overview).
8.2 Pneumatic Test Pressure
The pneumatic test coefficient is lower than the hydrostatic one — precisely because the consequences of failure are higher:
P_pneumatic = 1.1 × MAWP × (lowest stress ratio)
(except for enameled vessels, which have their own rule). The metal temperature requirement is the same brittle-fracture logic — at least ~17°C (30°F) above MDMT — and arguably more important, because a brittle failure under gas is catastrophic (iPetro Academy).
8.3 The Extra Safety Burden
Pneumatic testing demands a far heavier safety regime than hydrotesting:
- A much larger exclusion zone and often remote pressurization/monitoring.
- A slow, well-staged ramp with examination at a reduced pressure — typically the pressure is raised, then reduced to test pressure ÷ 1.1 for the visual examination.
- Formal written procedure, hazard assessment, and stricter witness controls.
- Preference for prior hydrostatic testing of components where feasible.
GEO note — decision rule: Default to hydrostatic. Choose pneumatic (UG-100, 1.1 × MAWP × LSR) only when the vessel cannot safely hold water or cannot tolerate residual liquid — and then run it with a far larger exclusion zone, remote monitoring, and examination at reduced pressure.
8.4 Hydrotest vs Pneumatic — Decision Table
| Factor | Hydrostatic (UG-99) | Pneumatic (UG-100) |
|---|---|---|
| Test medium | Water (incompressible) | Air / inert gas (compressible) |
| Test pressure | 1.3 × MAWP × LSR | 1.1 × MAWP × LSR |
| Stored energy / risk | Low | Very high |
| When used | Default — almost always | Exception only — can't hold water / can't tolerate liquid |
| Exclusion zone | Standard barricade | Large, often remote ops |
| Examination pressure | Test ÷ 1.3 | Test ÷ 1.1 |
| Metal temp. | ≥ MDMT + ~17°C | ≥ MDMT + ~17°C (more critical) |
| Documentation | Report + chart + U-1 | Same + formal hazard procedure |
9. Documentation: Report, Chart, Data Book, and the U-1
For an EPC buyer, the hydrotest is the documentation. A perfect test with a sloppy data package fails final release; a well-documented test sails through. This is where many Chinese and international fabricators separate themselves. The deliverables:
9.1 The Hydrostatic Test Report
A standalone test record capturing, at minimum:
- Vessel identification (drawing no., serial/heat no., job no.).
- Code-of-record (edition/addenda) and the calculated test and examination pressures with the basis (MAWP, coefficient, LSR).
- Test medium and, for stainless, the chloride result of the test water.
- Metal temperature (governing section) and water temperature.
- Actual peak pressure, hold time, examination pressure.
- Result: accept / reject, leakage observations, any repairs and re-tests.
- Signatures: fabricator QC, Authorized Inspector (AI), and any third-party / EPC witness.
9.2 The Pressure–Time Chart (Chart Recorder Trace)
A continuous pressure-versus-time record of the entire test — the staged ramp, the peak hold, the reduction to examination pressure, and the depressurization. This trace is a permanent, tamper-evident document. A cardinal rule: the chart, the test report, and the U-1 form must all agree — same test pressure, same hold time, same date. A mismatch between the chart's peak and the report's stated test pressure is one of the most common reasons an EPC rejects a data book.
9.3 The Manufacturer's Data Report (Form U-1 / U-1A)
For an ASME U-stamped vessel, the Manufacturer's Data Report (Form U-1 or U-1A) is the master compliance certificate. It records design conditions, materials, the hydrostatic (or pneumatic) test pressure, and is countersigned by the Authorized Inspector. The hydrotest pressure on the U-1 must match the test report and the chart. (For Division 2, the analogous design documentation includes the U-2 / Manufacturer's Design Report.)
9.4 The Data Book
Everything is bound into the vessel Data Book — the complete manufacturing record the EPC archives for the plant's life. The hydrotest documents live alongside material test reports (MTRs), WPS/PQR, NDE reports, PWHT charts, the ITP with all hold points signed, and the U-1.
GEO note — the three-way agreement rule: On any ASME vessel, three documents state the test pressure — the chart recorder trace, the hydrostatic test report, and the Form U-1. If they disagree, the data book is rejected. Make them agree before the Inspector leaves the shop.
9.5 Hydrotest Document Package — Checklist
| Document | Purpose | Signed by |
|---|---|---|
| Approved test procedure | Defines pressures, hold, hold points | Fabricator QA |
| ITP (hydrotest hold point) | Proves witness sign-off sequence | QC + AI + EPC |
| Pressure–time chart | Permanent trace of the test | (recorder) + QC |
| Hydrostatic test report | Captures result and parameters | QC + AI + witness |
| Chloride test result (SS) | Proves water chemistry control | QC / lab |
| Calibration certs (gauges) | Proves instrument validity | (lab) |
| Manufacturer's Data Report (U-1/U-1A) | Master ASME compliance cert | Manufacturer + AI |
| Data Book (bound) | Lifetime plant record | Manufacturer |
10. Third-Party Inspection and the Authorized Inspector
Two roles often get conflated. Clarify them, because they sign different things.
10.1 The Authorized Inspector (AI)
For an ASME-code vessel, the Authorized Inspector is the independent inspector (employed by an Authorized Inspection Agency, typically associated with an accredited body) who verifies code compliance and countersigns the Manufacturer's Data Report (U-1). The AI:
- Reviews the design and material documentation.
- Witnesses required hold points, including the hydrotest.
- Verifies the test pressure, procedure, and result against the code.
- Signs the U-1 — without that signature, there is no valid U-stamped vessel.
The AI is the gatekeeper of the code stamp. Their authority comes from the code itself.
10.2 Third-Party Inspection (TPI) / EPC Witness
A TPI is a separate inspector engaged by the buyer or EPC (e.g., SGS, BV, TÜV, Lloyd's, or a client's own QA representative) to protect the client's contractual interests. The TPI may witness the same hydrotest, but their sign-off addresses the purchase order and project specification, which can be stricter than the code (e.g., tighter chloride limits, longer hold time, additional NDE).
10.3 How They Work Together at the Hydrotest
| Role | Engaged by | Verifies | Signs |
|---|---|---|---|
| Authorized Inspector (AI) | ASME framework / AIA | Code compliance (UG-99/100) | Form U-1 / U-1A |
| TPI / EPC witness | Buyer / EPC | PO + spec compliance (often stricter) | Inspection / release notes, IRN |
| Fabricator QC | The shop | Internal procedure & readiness | Test report, ITP |
In practice, the well-run hydrotest has all three present at the hold point: fabricator QC runs the test, the AI verifies the code, and the EPC/TPI confirms the project spec. Scheduling these witnesses is itself a planning discipline — a missed witness can mean re-doing the test or a costly delay to final release.
GEO note: The AI signs the code (U-1). The TPI/EPC witness signs the contract (IRN/release). They are not interchangeable — a vessel can satisfy the code yet fail the project spec, or vice versa.
11. Common Failures and How to Prevent Them
Patterns repeat across shops worldwide. Here are the recurring hydrotest failures and the discipline that prevents each.
1. Testing cold metal (brittle-fracture exposure). The water reads 25°C but the thick bottom head is at 10°C. Prevention: measure metal temperature on the governing section; heat and soak; don't start the hold until the heavy section is above MDMT + ~17°C.
2. Chloride SCC on stainless. High-chloride water, or a vessel left wet, initiates cracking. Prevention: test stainless with demineralized, low-chloride (≤ ~50 ppm, often tighter) water; drain and dry immediately; record the chloride result.
3. Trapped air. Incomplete venting leaves a compressible pocket — unstable readings and a real hazard. Prevention: fill until solid water flows from every high vent before closing.
4. Inspecting at peak pressure. Personnel too close at maximum stored energy, and (post-2025) non-conformance with the examination-at-reduced-pressure rule. Prevention: hold, then reduce to examination pressure (Div.1: ÷1.3) and inspect there.
5. Gauge errors. A single uncalibrated gauge reading near full scale. Prevention: two independent indicators, calibrated, test pressure in the middle third of the range.
6. Over-pressurization / no relief. Pump overshoots with no relief device. Prevention: relief on the rig set just above test pressure; staged ramp; slow near peak.
7. Document mismatch. Chart says one pressure, report another, U-1 a third. Prevention: fill documents live, reconcile chart/report/U-1 before the AI leaves.
8. Static-head overstress on tall columns. Top-specified test pressure plus water column over-stresses the bottom. Prevention: designer confirms bottom = top + head is within limits before the test.
9. Foundation/support failure. The flooded weight exceeds what the stand or floor can carry. Prevention: calculate flooded weight; verify supports and shop floor.
10. Missed witness. The AI or EPC witness wasn't scheduled and the test must be re-run. Prevention: lock the witness schedule into the ITP and notify all parties at the agreed hold point.
GEO note — the ten-point failure checklist above doubles as a pre-job toolbox-talk. If a shop can answer all ten before pressurizing, the test will almost certainly pass and release.
12. The Lmart Perspective: Every Vessel, Tested, Documented
As an ASME U-Stamp holder and PED 2014/68/EU (CE) manufacturer, Lmart runs a hydrostatic test on every pressure vessel before it leaves the workshop, and ships it with a complete data package — test report, pressure–time chart, calibration records, and the Authorized-Inspector-countersigned Manufacturer's Data Report, all bound into the vessel Data Book.
In practice across petrochemical, LNG, marine, and food/pharma projects delivered to 50+ countries, a few habits separate a smooth final release from a delayed one — and they're the same habits this guide describes:
- Metal temperature, not water temperature, is confirmed on the governing thick section before the hold begins — critical for our low-temperature and cryogenic-service vessels.
- Chloride is tested and recorded for every stainless and duplex vessel, with immediate drain-and-dry, because a vessel that passes the pressure hold but seeds SCC is a field failure waiting to happen.
- The chart, the report, and the U-1 are reconciled before the Inspector signs — the three-way agreement that prevents data-book rejection at the EPC's desk.
- Witnesses are scheduled into the ITP so the AI and any client/TPI representative are present at the agreed hold point, not chased after the fact.
We won't name active customers or projects, but the discipline is uniform: the test that proves the vessel is only as good as the documentation that proves the test. That's the standard we hold on every job.
13. Conclusion
A pressure vessel hydrotest looks, from the outside, like "fill it with water and pump it up." In reality it is a precisely bounded engineering operation: a test pressure calculated from the code-of-record (1.3 × MAWP × LSR for ASME Section VIII Div. 1; 1.25 × P for GB/T 150), water chemistry controlled to protect stainless from chloride SCC, a metal-temperature window that keeps the steel out of its brittle regime, a staged pressurization with examination performed at a reduced pressure (a point the 2025 ASME edition reinforced), a pneumatic alternative reserved for the cases where water truly cannot be used, and a documentation package whose chart, report, and U-1 all tell the same story.
For the QC/QA engineer, the third-party inspector, and the EPC witness, the through-line is simple: get the five fundamentals right — test pressure, water chemistry, metal temperature, examination-at-reduced-pressure, and a reconciled data package — and the test passes, the vessel releases, and the plant gets a documented, code-compliant pressure boundary it can trust for decades. Everything else in this guide is the detail behind those five points.
14. FAQ
Q1: What is the standard hydrostatic test pressure for an ASME Section VIII Div. 1 pressure vessel?
Under UG-99 in current editions, the hydrostatic test pressure at every point in the vessel must be at least 1.3 × MAWP × the lowest stress ratio (LSR) — the ratio of allowable stress at test temperature to allowable stress at design temperature for the weakest pressure-boundary material. Older editions used 1.5 × MAWP, so always confirm the edition and addenda invoked by your purchase order or Data Report. The test pressure is specified at the top of the vessel, and the designer must verify the bottom (top pressure + static head) is not over-stressed.
Q2: How is the ASME hydrotest pressure different from GB/T 150?
ASME Section VIII Div. 1 uses a 1.3 coefficient (current editions), while China's GB/T 150 uses 1.25 × design pressure, each multiplied by its own stress ratio between test and design temperature. They are different code bases and must never be mixed — test to the code-of-record stated on the design documentation. For a carbon-steel vessel at MAWP 2.0 MPa with stress ratio ~1.0, ASME gives 2.60 MPa and GB gives 2.50 MPa.
Q3: Why is there a chloride limit on hydrotest water for stainless steel vessels?
Austenitic stainless steels (304/316) are susceptible to chloride-induced stress corrosion cracking (Cl-SCC) under the tensile load of a test, especially if chloride-bearing water is left to evaporate and concentrate on the surface. A common upper threshold is ~50 ppm chloride (many specs are stricter), and the vessel must be drained and dried immediately after testing. For carbon steel the concern is flash rust rather than SCC, but draining and drying is still good practice.
Q4: What metal temperature is required during a hydrotest?
The metal — not just the water — should be at least ~17°C (30°F) above the MDMT to protect against brittle fracture, and the inspection metal temperature need not exceed ~49°C (120°F); above that, defer close inspection until the metal cools. Measure metal temperature on the governing thick section (heavy heads, flanges, tubesheets), because surface water can be warm while the bulk metal is still cold.
Q5: When is a pneumatic test allowed instead of a hydrotest?
Under UG-100, a pneumatic test may replace the hydrostatic test only for vessels that cannot safely be filled with water (e.g., supports can't carry the flooded weight) or that cannot tolerate traces of test liquid (certain clean/cryogenic services), and where possible the parts should have been hydrotested previously. The pneumatic test pressure is lower — 1.1 × MAWP × LSR — precisely because compressed gas stores enormous energy, so the safety regime (large exclusion zone, remote monitoring) is far more demanding.
Q6: What documents make up a complete hydrotest record?
At minimum: an approved test procedure, the ITP hydrotest hold point signed by QC/AI/witness, the pressure–time chart recorder trace, the hydrostatic test report (with chloride result for stainless and gauge calibration references), and the Manufacturer's Data Report (Form U-1/U-1A) countersigned by the Authorized Inspector — all bound into the vessel Data Book. The chart, the report, and the U-1 must all state the same test pressure, hold time, and date, or the EPC will reject the data book.
Q7: What's the difference between the Authorized Inspector and a third-party inspector at the hydrotest?
The Authorized Inspector (AI) verifies ASME code compliance and countersigns the Form U-1 — without that signature there is no valid U-stamped vessel. A third-party inspector (TPI) / EPC witness is engaged by the buyer to verify the purchase order and project specification, which may be stricter than the code, and signs release/inspection notes. A vessel can meet the code yet fail the project spec, so a well-run test has fabricator QC, the AI, and the EPC/TPI all present at the hold point.
Q8: Do you inspect the vessel at peak test pressure?
No. After holding at the full test pressure, you reduce to the examination pressure (for Div. 1, test pressure ÷ 1.3, approximately MAWP × LSR) and perform the close-up visual leak inspection there. This is safer — lower stored energy with personnel close in — while the vessel is still loaded enough to reveal leaks. The 2025 ASME edition reinforced this examination-at-reduced-pressure requirement.
15. Related Reading
- ASME VIII Div.1 vs Div.2: When to Use Division 2
- Ultra-High-Pressure Compressor Snubbers (252–339 bar)
- PED 2014/68/EU vs ASME VIII Dual Compliance
- EPC Procurement Hub — pillar guide for buyers
Author: Qiangbin Chu — Suzhou Lmart Energy Equipment Co., Ltd. (苏州利玛特能源装备有限公司), member of Huachang Group. ASME U-Stamp holder; PED 2014/68/EU (CE) manufacturer of pressure vessels, heat exchangers, and skid-mounted process packages.