How Do EPC Buyers Review Factory Acceptance Testing for Large Pressure Vessels and Heat Exchangers?

Missing a hold point at hydrostatic testing is not a paperwork problem — it voids your right to dispute a weld defect found six months into commissioning, on a vessel now buried inside a heat recovery train. EPC buyers who treat factory acceptance testing as a box-checking formality routinely absorb the consequences downstream: re-mobilization of third-party inspectors, schedule compression on interconnecting piping, and — in the worst cases — pressure test failures on site where regulatory consequences dwarf anything that would have happened inside the fabricator’s shop.

EPC buyers review factory acceptance testing for pressure vessels and heat exchangers by auditing the Inspection and Test Plan against the applicable code (typically ASME VIII Div.1 or equivalent), witnessing mandatory hold points — radiographic examination, hydrostatic test at 1.3× MAWP, and nameplate verification for vessels; bundle pull and dual-side pressure tests at 1.25–1.5× design pressure for heat exchangers — and confirming all certified documentation is traceable before signing the release note.

What surprises most procurement managers who have not sat through a real FAT is how much leverage disappears the moment a release note is signed without an open punch list. The fabricator’s contractual obligation compresses sharply after that signature. Understanding exactly which hold points carry legal and insurance weight, how far in advance an ITP must land on the EPC’s desk, and what a credible witness inspection schedule actually looks like — that sequence is what separates a smooth equipment delivery from a claim waiting to happen.

Decoding the Inspection and Test Plan: What EPC Reviewers Look for Before FAT Begins

The ITP is not a formality. It is the contractual backbone that determines whether the FAT will catch real problems or simply rubber-stamp a vessel that should have failed. Experienced EPC QA engineers read an ITP the way a structural engineer reads a weld map — looking for what is missing as much as what is there.

ITP Structure: The Five Columns That Actually Matter

A properly constructed ITP is organized in tabular form, with each inspection activity broken into at minimum five defined fields: the activity description, the governing code or standard reference, the quantified acceptance criteria, the responsible party (vendor, third-party inspector, or owner’s engineer), and the hold/witness/review point designation. The H/W/R system is where enforceability lives. A Hold point means fabrication cannot proceed without the designated party’s physical sign-off. A Witness point means the party must be notified and given the opportunity to attend. A Review point means documentation is submitted for desk review after the fact.

EPC buyers flag ITPs that over-classify activities as Review when the risk profile demands a Hold. Hydrostatic testing at 1.3× MAWP under ASME Section VIII Div.1 is never a Review point on a critical service vessel — it should be a Hold for the owner’s inspector, full stop. Similarly, final nameplate verification and MAWP stamping are Hold points, not witness events.

Cross-Checking the ITP Against Purchase Order Documents

Before any fabrication begins, the EPC QA team lays the draft ITP alongside the purchase order data sheet, the vessel design specification, and the referenced codes — typically ASME VIII Div.1 or Div.2, TEMA class R, C, or B for heat exchangers, and applicable API standards such as API 661 for air-cooled exchangers. The goal is to identify three categories of mismatch: activities that appear in the spec but are absent from the ITP, acceptance criteria in the ITP that are looser than the code or spec allows, and code references that are listed but not edition-specific.

An ITP that cites “ASME IX” for welder qualification without specifying the applicable welding procedure specification numbers tells the reviewer nothing enforceable. The same applies to NDE coverage: stating “RT per ASME V” without defining the film density range, the penetrameter type, or the percentage of joint coverage leaves the acceptance threshold undefined until someone disputes a film.

Common Deficiencies Experienced QA Engineers Flag Immediately

Missing NDE method statements are the most frequent gap. RT, UT, MT, and PT each require a written procedure defining equipment calibration, surface preparation, technique, and acceptance standard. If the ITP lists “PT on nozzle welds” without referencing a documented procedure per ASME V Article 6, the activity has no enforceable baseline.

Undefined PWHT soak temperature tolerances are a close second. Specifying “PWHT per UCS-56” is not sufficient on its own. The ITP should state the soak band — typically ±14°C around the target temperature for carbon steel, though the actual range depends on material P-number group and section thickness — along with thermocouple placement requirements and chart recorder calibration intervals.

Hydrostatic test fluid requirements — water temperature minimums to avoid brittle fracture risk, maximum chloride content for stainless or duplex vessels, and post-test drying or nitrogen blanketing for moisture-sensitive service — are routinely absent from vendor-drafted ITPs and routinely demanded by EPC reviewers once caught.

ITPs for Complete Evaporative Condensation Systems: Beyond the Main Shell

This is where ITPs for integrated systems expose their weakest seams. A vessel-level ITP that covers only the main shell and its nozzles is inadequate for a complete evaporative condensation system. The ITP must address steam-side nozzle flange facing inspection — surface finish Ra values and gasket seating surface flatness matter when the flanges will see thermal cycling — and confirm compatibility with the specified gasket type.

Cooling water inlet strainer verification must appear as a discrete line item, including mesh size confirmation, pressure drop across the clean strainer at design flow, and a visual check for manufacturing debris retained in the basket. Vacuum system interface integrity requires a documented hold-down vacuum leak test on the vapor-side connections, not just a shell-side hydrostatic. Condensate return line cleanliness — often the last thing fabricators clean before shipment — needs a defined flushing procedure and acceptance criterion for particulate count or visual cleanliness, depending on how sensitive the downstream condensate recovery system is.

An ITP that treats these subsystem interfaces as out-of-scope is one that will generate field punch list items during commissioning, which is where rework costs four to ten times what it costs at the shop.

Material Traceability and Mill Certificate Review: The First Hard Gate in FAT

Before a single weld bead is laid, before the hydrostatic pump is connected, the EPC QA engineer is already looking for gaps — and the most common place those gaps appear is in material documentation. For large pressure vessels and heat exchangers, material traceability is not a paperwork formality. A broken chain between the mill certificate, the plate marking, and the fabrication traveler is enough to put a vessel on hold, sometimes weeks before it was supposed to ship.

Heat Number Traceability: What “Complete” Actually Means

Every plate, pipe, and forging on an ASME VIII Div.1 vessel carries a heat number rolled or stenciled into the material. The material test report (MTR) — issued by the steel mill — records the chemical composition and mechanical properties for that specific heat. The fabricator’s job is to preserve that chain: the heat number on the MTR must match the marking on the physical plate, which must appear on the cutting plan, the weld map, and ultimately the fabrication traveler that follows each component through the shop.

EPC reviewers typically audit this chain at three points: when the MTR package is submitted (usually 2–4 weeks before the first hold point), during a pre-FAT document review, and physically on the shop floor by cross-checking plate remnants against the traveler. If a plate has been cut and the heat number was not transferred to the sub-piece — a genuinely common shop-floor lapse — the entire component is quarantined until traceability is restored or the material is re-identified through PMI testing.

PMI Testing: When 100% Coverage Is Non-Negotiable

Positive material identification with an XRF analyzer closes the gap when visual markings are ambiguous or absent. The scope of PMI varies significantly depending on the material and the client’s specification. For wetted parts fabricated from duplex stainless steel (such as UNS S31803 or S32205), Hastelloy alloys, or titanium, many EPC specifications — particularly those following IOGP S-563 or equivalent company standards — require 100% PMI on all wetted components, including nozzles, weld overlay surfaces, and heat exchanger tubes. Carbon steel vessels on non-corrosive service are more commonly covered by spot-check protocols, typically 10–20% of components, though this range depends on the owner’s risk tolerance and the specific service.

Charpy Impact Testing for Low-Temperature Service

Vessels designed for low-temperature service — cryogenic exchangers, condensers operating below −20 °C, separators in cold-climate installations — require impact-tested materials. The EPC buyer’s engineer will look for Charpy V-notch results at the design minimum temperature (DMT) on the MTR itself, not in a separate supplemental report. If the MTR only states “impact tested” without listing the test temperature and absorbed energy values in joules or foot-pounds, it fails the review.

The vessel data sheet specifies the DMT and the minimum absorbed energy threshold. Those two values must be traceable directly to the MTR. A mismatch — for example, material tested at −29 °C when the DMT is −40 °C — triggers a material rejection regardless of how far fabrication has progressed.

Consumable Traceability: The Overlooked Audit Point

Welding consumables carry their own traceability requirements, and this is where fabricators most frequently come unprepared. Each spool of filler wire, each batch of flux, and each shielding gas cylinder used on a pressure-retaining weld must be documented with a heat or lot number that links back to a manufacturer’s certificate of conformance. The certificate must confirm the consumable meets the applicable AWS or equivalent specification.

Stringent EPC QA engineers will walk the welding stations and ask to see the consumable log alongside the weld traveler for each joint. If a welder pulled wire from an unmarked spool or a flux drum without a legible lot number, that weld joint enters a grey zone. Depending on the client’s specification, resolution may require additional NDE — adding cost and schedule time the fabricator did not budget.

Auxiliary Vessels and Separate MTR Packages

In a condensation system, the primary exchanger or vessel rarely ships alone. Condensate collection tanks, waste liquid storage tanks, and vacuum receiver vessels are typically fabricated from different material grades — often carbon steel where the main vessel is alloy — and each requires its own MTR package organized and indexed separately.

EPC buyers conducting a joint FAT expect to receive a complete document package for every vessel in the scope, not just the highest-pressure item. A fabricator who delivers a polished documentation set for the main exchanger but hands over loose, un-indexed certificates for the condensate tank will see the entire FAT paused while that package is reconstructed. The practical lesson: treat auxiliary vessel documentation with the same discipline as the primary vessel, and organize all MTR packages into a single indexed binder — one section per vessel, cross-referenced to the equipment tag on the ITP — before the EPC review team arrives.

Weld Inspection Hold Points: NDE Witness Sequence for Shells, Nozzles, and Tube Sheets

EPC buyers don’t treat NDE as a box-ticking exercise at the end of fabrication. They build a sequential hold-point structure into the ITP precisely because catching a rejectable indication after PWHT is exponentially more expensive — in time, repair heat cycles, and schedule — than catching it before the shell rolls into the furnace. Understanding that sequence, and why each step gates the next, is what separates an experienced FAT witness from someone reading a checklist.

The Sequential Hold Point Logic

Fabrication witness typically follows this order: visual weld inspection (VT) on each weld pass or at minimum on the root and cap, followed by a dimensional check against the approved fabrication drawing, then volumetric examination — RT or UT — on longitudinal and circumferential seams, and finally surface examination (MT or PT) on nozzle attachment welds and any full-penetration structural attachments. Each step must be formally signed off before the next begins. A fabricator who shoots radiographs before completing dimensional verification, or who performs PT on nozzle welds before the dimensional check is closed, gives the EPC buyer grounds to reject the entire NDE package and require restart — which has happened on projects where the shop was rushing toward a pressure test window.

RT Versus UT: How EPC Buyers Make the Call

The choice between radiographic and ultrasonic examination is not purely technical — it carries code-credit implications that directly affect the design. Under ASME Section VIII Div.1, a Type 1 butt weld with 100% RT qualifies for a joint efficiency of 1.0; spot RT drops that to 0.85. EPC buyers reviewing the design basis will verify which credit the vessel was designed to, then confirm the ITP mandates the corresponding examination scope. For wall thicknesses above roughly 50–75 mm (the range depends on the specific weld geometry and RT sensitivity requirements), phased-array UT often becomes the preferred or even required method because film RT loses sensitivity on thick sections. ASME Code Cases such as Code Case 2235 allow UT in lieu of RT, but only when the procedure and personnel qualifications are pre-approved by the authorized inspector — the EPC buyer’s witness will ask to see those procedure qualifications before the first scan begins, not after.

Tube-to-Tube-Sheet Welds and Eddy Current Baseline

Heat exchanger tube-to-tube-sheet joints get a separate, more granular inspection sequence. The distinction between strength welds and seal welds is not semantic: strength welds carry mechanical load and require full WPS/PQR qualification, typically with cross-section macro examination on qualification coupons. Seal welds, common on lower-pressure units, still require PT or MT after welding. Before bundle assembly, EPC buyers increasingly mandate an eddy current baseline scan of the tubes — particularly on high-alloy or clad tube materials — to establish a pre-service reference that can be compared against in-service inspection data years later. The tube expansion qualification procedure (roller expansion, explosive, or hydraulic) must also be witnessed: pull-out load testing on a sample set, with pass/fail criteria defined in the client specification rather than left to the fabricator’s judgment.

PWHT Documentation: The Chart Recorder Problem

PWHT chart recorder traces are a chronic source of rejection at FAT. The common fabricator error is submitting recorder charts that show the temperature profile but lack a current thermocouple calibration certificate traceable to a national standard, or charts where the thermocouple count doesn’t match the number specified in the heat treatment procedure — a vessel with a long shell may require six or more thermocouples to verify uniform soak temperature. Hardness test results on both the weld and the heat-affected zone must appear in the certified report; results reported only on the base metal tell the EPC reviewer nothing about whether the HAZ was adequately tempered.

Nozzle Orientation and Flange Face Finish as a Final Pre-Test Gate

Once all weld NDE and PWHT documentation are closed, nozzle orientation — verified against the certified fabrication drawing in degrees from a defined datum — becomes a formal hold point before hydrostatic test authorization is granted. A nozzle rotated even a few degrees can conflict with piping take-offs in a congested equipment bay, and correcting that in the field costs far more than a re-check at the shop. Flange face finish is measured with a profilometer; most process flanges require 125–250 AARH (arithmetic average roughness), with tighter finishes specified for spiral-wound gaskets on high-pressure or hydrogen service. EPC buyers who skip this check routinely encounter gasket seating failures during initial pressurization on site — a consequence that traces directly back to a missed hold point during FAT.

Hydrostatic and Pneumatic Pressure Testing: Procedures, Safety Protocols, and Acceptance Criteria EPC Buyers Enforce

Pressure testing is the moment when months of fabrication either hold or fail. For EPC buyers witnessing FAT, this is not a formality — it is the primary mechanical integrity verification they will defend to their client if something goes wrong in the field.

Test Pressure Calculation: Why the Stress Ratio Actually Matters

ASME VIII Division 1 specifies hydrostatic test pressure as 1.3 × MAWP × (S_test / S_design), where S_test and S_design are the allowable stresses at test temperature and design temperature respectively. Fabricators routinely quote 1.3× and stop there. EPC buyers who know the code check the stress ratio.

For vessels with austenitic stainless steel shells operating at elevated design temperatures — say 300–400 °C — the allowable stress at ambient test temperature is measurably higher than at operating temperature. That ratio can push the actual required test pressure 8–15% above a naive 1.3× calculation, depending on material grade and the temperature spread. An EPC reviewer will pull UG-99(b) and cross-check the fabricator’s test pressure against the certified material stress tables in Section II Part D. If the fabricator used a nominal multiplier without applying the ratio, that is a hold-point failure before the first drop of water enters the vessel.

Test Fluid Requirements and Brittle Fracture Risk

Water for hydrostatic testing must be clean — typically potable or demineralized — and temperature-controlled. ASME VIII Div.1 UG-99 requires water temperature to be at least 17 °C (30 °F) above the vessel’s minimum design metal temperature (MDMT) during pressurization. For low-alloy carbon steel vessels with an MDMT of −20 °C, that means test water should be at or above roughly 0–3 °C as a minimum. In practice EPC specs often require 15–25 °C to provide a comfortable margin against brittle fracture — particularly relevant in winter fabrication shops where large vessels sitting on a cold shop floor can drop significantly in local metal temperature.

For austenitic stainless steel vessels, chloride content in the test water is a hard limit. Most EPC specifications set this at below 25 ppm, and some clients in the chemical and pharmaceutical sectors specify below 10 ppm. This is not conservative paperwork. Chloride-induced stress corrosion cracking can initiate under a hydrostatic test on a sensitized 304 or 316L weld if the water sits inside the vessel for days without being properly drained and dried.

Gauge Calibration and Placement Requirements

Two calibrated pressure gauges are required: one at the highest point of the vessel to monitor the minimum local pressure, one at the test pump connection or vessel bottom. Calibration certificates must be current — typically within six months — and EPC buyers will physically check the certificate dates on-site before pressurization begins.

The gauge range requirement is operationally significant. Test pressure should fall between 25% and 75% of gauge full scale. A 600 psi gauge on a 100 psi test is useless for detecting a slow leak. An EPC reviewer who finds incorrect gauge ranging will stop the test and require substitution. This happens more often than fabricators expect, particularly on smaller vessels where technicians grab whatever gauge is available.

Pneumatic Testing: Permitted Cases and the Safety Burden

Pneumatic testing is permitted under ASME VIII Div.1 UG-100 when vessels cannot be safely filled with liquid — internal linings that would be damaged, structural load limitations, or process systems that cannot tolerate any moisture. EPC buyers treat pneumatic test requests with scrutiny because the stored energy in a compressed-gas test is orders of magnitude higher than in an equivalent hydrostatic test. Approval requires a documented technical justification, not just a convenience argument.

The code-mandated sequence includes a preliminary pressurization to 25 psi above MAWP (or 10% of test pressure, whichever is lower), a hold and inspection, then incremental steps to full test pressure. EPC buyers typically require that the safety exclusion zone radius be calculated and physically barricaded before any pressurization begins. Personnel count inside the zone is restricted. Time-stamped photographic documentation of the barricade setup is a standard EPC requirement, and the pneumatic test procedure itself must be reviewed and signed off before the test date.

Post-Test Walkdown and Documentation Before Bleeddown

At test pressure, pressurization stops and the physical walkdown begins. The EPC buyer’s inspector and the third-party inspector move through every accessible weld seam, nozzle attachment weld, and flange face looking for weeping, distortion, or audible indication. No pressure bleeddown is authorized until the walkdown is complete and documented.

Time-stamped photographs — typically taken by TPI with the EPC buyer’s representative visible in frame or confirmed as present — record the vessel at test pressure with the gauge reading legible. This documentation travels with the vessel’s material dossier to site and can be reviewed during commissioning or an insurance audit years later.

Heat Exchanger Shell-Side and Tube-Side Sequential Testing

Heat exchangers require separate pressure tests on each side, and the sequence matters. Shell-side and tube-side design pressures often differ significantly. Where the design basis includes one side under full pressure while the other operates under vacuum — common in condenser and reboiler service — EPC buyers verify that the test procedure explicitly accounts for the differential pressure case.

A tube bundle with tube-to-tubesheet joints strong enough to pass a 1.3× shell-side hydrotest may still leak at the joint root under the combined mechanical loading of internal pressure on one side and external vacuum on the other. EPC reviewers check that the fabricator’s ITP addresses this scenario, not just the straightforward one-side-at-a-time test that passes under benign conditions.

Functional and Dimensional Acceptance: Nozzle Orientation, Flange Mating, and System Interface Verification for Complete Condensation Skids

Pressure testing gets the attention, but dimensional and interface verification is where fabrication errors that actually delay commissioning tend to surface. For an integrated evaporation-condensation package — where a process vessel, shell-and-tube condenser, condensate collection tank, waste liquid storage tank, and vacuum system all ship as a coordinated skid or matched set — a single mislocated nozzle discovered at site can cost more in remediation than several weeks of additional FAT time would have.

Certified Dimensional Report: What the Package Must Contain

EPC buyers require a certified dimensional inspection report signed by the fabricator’s QC inspector and, at witness hold points, countersigned by the third-party inspector. The report maps measured values against the approved general arrangement (GA) drawing, with explicit tolerances called out for each dimension. Key parameters include overall vessel length or height (typically ±3–6 mm on vessels over 3 m, depending on specification), nozzle center-to-center spacing, nozzle projection from shell outside diameter, and vessel centerline levelness measured at saddle or skirt base plates.

Nozzle projection matters because field piping spools are pre-fabricated against the certified drawing dimensions. If the actual projection deviates by even 10–15 mm beyond tolerance, the field spool either cannot be made up or requires a field modification — which on an EPC project running parallel construction activities is rarely a simple fix.

Flange Bolt-Hole Orientation: A Small Detail With Large Consequences

Every process nozzle flange has a bolt-hole circle, and on flanges above DN 150 / 6 inches the bolt holes straddle the vertical or horizontal centerline unless the drawing specifies otherwise. A single nozzle rotated 22.5 degrees — one bolt-hole pitch on a standard eight-bolt pattern — means the field piping flange cannot be made up without either shimming the pipe spool or reorienting the nozzle. Reorienting a nozzle on an already-hydrostatically-tested vessel means cutting, re-welding, re-PWHT where applicable, and re-testing. At the fabricator’s shop this might take one to two weeks. At site, factor in scaffolding, confined space permits, and schedule interference with other trades — the real cost is three to five times higher, and that is before any liquidated damages exposure.

EPC buyers typically verify bolt-hole orientation using a simple template or digital angle measurement against the orientation callout on the nozzle schedule. This takes minutes at FAT. Skipping it is not conservative — it is a known source of expensive field rework.

System Interface Verification for Evaporation-Condensation Packages

When the FAT scope covers a complete condensation package rather than a single vessel, the interface verification becomes a system-level exercise. The steam supply nozzle on the process side must match the specified pressure class and facing — raised face, ring joint, or flat face — because the upstream steam control valve assembly is often pre-purchased separately and arrives with a fixed end connection. Cooling water inlet and outlet nozzle locations must correspond to the tube bundle orientation: on a horizontal shell-and-tube condenser, the inlet is typically at the bottom of the water box and the outlet at the top to promote full tube-side flooding, and any deviation from this arrangement on the fabricated unit needs engineering sign-off before FAT closes.

Vacuum system connection nozzles require particular attention. The nozzle rating, the blank-off provision for pressure testing (since vacuum equipment is often not present at FAT), and the connection elevation all affect whether the installed vacuum pump can achieve and maintain design suction pressure. A nozzle positioned 200–300 mm higher than the P&ID elevation — a common drafting-to-fabrication slip — changes the static head the vacuum pump works against and can shift system performance outside the vendor’s guaranteed curve.

Condensate return nozzle elevation must be verified against the overflow level of the condensate collection tank. If the condenser condensate outlet nozzle elevation is lower than the tank overflow nozzle elevation as built, condensate backs up into the condenser shell, reducing effective heat transfer area. This is a steady-state performance defect that will not show up during hydrostatic testing but will show up immediately during commissioning, and it is not always obvious what is causing reduced capacity.

Auxiliary Vessel Interface Checks Within the Same FAT

A disciplined EPC buyer extends interface verification to every vessel in the package during the same FAT visit rather than scheduling separate inspections. Condensate collection tank level instrument nozzles — for float switches, guided-wave radar, or differential pressure transmitters — must align with the nozzle schedule on the instrument data sheets, with the correct flange rating and correct insertion orientation (vertical, angled, or horizontal). Waste liquid storage tank vent and drain connections must match the P&ID both in size and in the direction of discharge, since vent orientation affects whether a conservation vent or flame arrester can be fitted without field modification. Vacuum pump seal water supply and return connections must be verified for size and elevation against the pump vendor’s installation drawing.

Instrument and Relief Device Mounting Verification

Thermowell insertion lengths must be verified against the process datasheet — a thermowell that bottoms out against the vessel wall rather than sitting at the specified immersion depth gives a skewed temperature reading. Pressure gauge root valve orientation matters for drainage on outdoor installations; a root valve oriented horizontally on a vertical nozzle traps condensate and can freeze in cold climates or contaminate gauge internals in dirty service.

Safety valve outlet direction is a code and platform-steel coordination issue. ASME and most international standards require that safety valve discharge does not endanger personnel or obstruct access routes, but the specific outlet orientation must also clear the platform structural steel shown on the skid layout drawing. Verifying this at FAT — before the platform is installed at site — takes a drawing overlay and a few minutes of physical inspection. Finding an interference at site means either modifying the relief valve discharge piping or modifying the platform steel.

Nameplate data cross-check is the final gate: every vessel nameplate MAWP, design temperature, shell material designation, and hydrostatic test pressure is compared against the vessel data sheet and, for pressure vessels under ASME jurisdiction, against the Manufacturer’s Data Report (MDR). Discrepancies here are not administrative — they affect the legal status of the vessel under the applicable pressure equipment regulation.

Documentation Package Handover: What EPC Buyers Audit Before Releasing Vessels for Shipment

A pressure vessel that has physically passed its hydrostatic test is not a releasable vessel. That distinction matters more than most fabricators appreciate until they are staring at a shipment hold two days before the scheduled truck departure. EPC buyers treat the manufacturing data record — the MDR, or quality dossier, depending on the contractor’s terminology — as a parallel deliverable of equal standing to the hardware itself. An incomplete package stops the clock regardless of what the pressure gauge showed.

Core MDR Contents and the Systematic Audit

The checklist an EPC lead inspector works through is not arbitrary. It maps directly to the code of record and the purchase order data requirements list. At minimum, buyers expect to verify: design calculations stamped by the engineer of record, approved-for-construction drawings with revision history, full material traceability records tied to each pressure-retaining component, welder qualification records (WPS, PQR, and WPQ per ASME Section IX), NDE reports with current interpreter certifications, post-weld heat treatment strip charts with temperature and time axes traceable to calibrated recorder serial numbers, the hydrostatic test record signed by the responsible inspector, and the final inspection certificate. Any single missing item is grounds to open a non-conformance. Buyers typically flag two categories of problems: outright missing documents, and documents present but unsigned or undated — the latter is surprisingly common for PWHT charts on carbon steel vessels where the furnace operator hands over raw strip chart rolls without a formal signatory.

ASME Code Stamp and Authorized Inspector Sign-Off

For ASME Section VIII Div.1 or Div.2 vessels, the U-1 or U-1A manufacturer’s data report must carry the Authorized Inspector’s original signature before the vessel physically leaves the shop floor. EPC buyers do not treat a verbal confirmation or an emailed scan as sufficient — they verify the AI’s commission number against National Board records and confirm the AI’s authorization covers the applicable code edition. This step matters because AI commission numbers expire and some smaller shops work with an AI whose commission has lapsed since the last major job. The National Board’s online registry makes this a 60-second check, but it is one that catches real problems several times a year across the industry.

As-Built Drawing Cross-Check and MDR Addenda

Fabrication never proceeds exactly as designed. Nozzle reinforcement pad sizes get revised, flange ratings get upgraded when plate stock changes, and branch connections shift axially to clear interference with structural steel. Each of those changes should travel through a formal field change notice and appear as a redline on the as-built drawing set. EPC buyers physically walk the vessel — or direct their third-party inspector to do so — comparing the nozzle schedule on the approved drawing against actual nozzle centerline elevations and orientations as measured on the vessel. For condensation skids where multiple nozzles interface with piping spooling fabricated off-site, a 10 mm nozzle elevation error discovered at site is a schedule and cost problem of a completely different magnitude than one caught during MDR review in the shop.

Coating and Surface Treatment Records

Surface treatment documentation is underestimated at the procurement stage and becomes contentious at handover. EPC buyers want blast profile records confirming Sa 2.5 or Sa 3 per ISO 8501-1, DFT logs for each individual coat with both wet film and dry film readings recorded by location, holiday test records for any immersion-service or buried-service coating system, and — for stainless and duplex vessels — pickling or passivation certificates with the chemical bath composition, temperature, and exposure time recorded. A DFT log that says “average 250 µm” without location-by-location readings will not satisfy a serious EPC buyer. The figure that matters is the minimum reading, not the average.

Preservation, Packing, and Shipping Document Review

The final audit layer covers how the vessel will travel and survive until commissioning. For stainless heat exchangers with nitrogen blanketing specified, buyers verify the recorded blanketing pressure and the valve or plug arrangement used to maintain it. Desiccant quantities inside flanged closures are checked against the purchase order preservation spec — quantities depend on enclosed volume and expected transit time, typically ranging from 1 to 4 moisture-absorbing units per enclosed cavity for standard sea freight durations. Temporary transport saddle drawings must be present and show the contact arc and timber specification to prevent shell distortion during road transit. The export packing list must align line-for-line with the purchase order. A crate labeled with a generic description that does not match the tagged equipment number on the PO creates customs, insurance, and receiving problems that cascade well beyond the paperwork itself.

Frequently Asked Questions About EPC Factory Acceptance Testing for Pressure Vessels and Heat Exchangers

What is the difference between a hold point and a witness point in a FAT ITP — and can a fabricator proceed past a hold point without the EPC buyer present?

No. A hold point is a hard stop written into the ITP that legally requires an authorized signature from the EPC buyer’s representative — and in most cases the Authorized Inspector (AI) as well — before fabrication or testing can continue. No signature, no forward progress. It is not a scheduling suggestion. Hydrostatic testing and final dimensional acceptance are the most common hold points on pressure vessel ITPs.

A witness point is different in practice. The fabricator must issue advance notification — typically 3–6 working days, as specified in the ITP — and the buyer has the right to attend. If the buyer or TPI representative does not show after reasonable notice, work can proceed and the activity is documented as witnessed-not-attended. That distinction matters enormously for schedule: a missed witness point rarely causes delay; a missed hold point always does. Fabricators who conflate the two, or who pressure buyers to convert hold points to witness points during schedule crunches, are exhibiting a procurement red flag worth noting in bid evaluation.

How many calendar days should EPC buyers budget for FAT of a large shell-and-tube heat exchanger above 100 metric tons?

Plan on 3–7 working days for the physical inspection sequence itself — this covers bundle pull observation, shell-side and tube-side pressure testing, dimensional checks, nozzle orientation verification, and nameplate review. The range depends on TEMA class, whether PWHT is required and complete, NDE backlog at the shop, and how many hold points are stacked in the ITP. Add 1–2 weeks minimum for documentation review: the AI sign-off on the data book, final MDR compilation, and any outstanding NCR closures. If radiographic film is still pending review or a PWHT chart is under dispute, documentation close-out can stretch beyond that. EPC teams that book flights assuming a three-day FAT on a complex unit frequently find themselves either extending the trip or accepting an incomplete sign-off — neither is a good outcome.

Can a fabricator substitute pneumatic testing for hydrostatic testing if water contamination is a process concern?

Yes, but the bar for approval is genuinely high. ASME VIII UG-100 permits pneumatic testing under specific conditions, and the fabricator must provide a code-compliant written justification — not just a verbal statement that “water is incompatible with the service.” The justification needs to address why hydrostatic testing cannot be done even with a subsequent dry-out procedure. Both the AI and the EPC buyer’s QA function must approve the substitution in writing before any test setup begins.

Expect the EPC buyer to require a formal risk assessment, a pressure relief valve on the test circuit sized for the gas supply, a preliminary leak check at 10% of test pressure, and a clearly defined personnel exclusion zone. Shops that treat pneumatic testing as a routine swap are shops that have not done it under serious EPC oversight before.

What happens when a pressure vessel fails the hydrostatic test at FAT — does the EPC contract automatically terminate?

It does not. Failure triggers a formal Non-Conformance Report (NCR) issued within 24–48 hours, typically by the AI or the TPI. The fabricator must then submit a root cause analysis and a proposed repair procedure — weld repair being the most common path — which both the AI and EPC QA must review and approve before any repair work begins. After repair, all affected NDE and the full hydrostatic test repeat. The documentation package is updated to include the original failure record, the NCR, the repair procedure, and the re-test results. The shipment hold stays in place until every one of those elements is resolved and signed off. Timelines for this process range from one week to over a month, depending on the defect type, the extent of repair, and whether heat treatment is required afterward.

For an integrated evaporation-condensation system, can auxiliary vessels — condensate tanks, waste liquid tanks, vacuum seal pots — be FAT-tested separately from the main evaporator?

Yes, and separate ITPs with individual test records for each vessel are standard practice. Each vessel is a discrete pressure boundary and gets its own hydrostatic test documentation. What the EPC buyer typically adds on top of that is a system-level interface verification step: once all vessels and skid components are assembled — or at minimum positioned and aligned per the General Arrangement drawing — the buyer’s representative performs a final check of nozzle orientations, flange face conditions, utility tie-in locations, and instrument connection alignment across the complete assembly. Condensate return lines, vacuum pull points, steam supply connections, and waste liquid drain interfaces all need to match the piping isometrics exactly. Discovering a 90-degree nozzle rotation error on a condensate tank after individual FAT sign-off is not uncommon, and catching it before shipment rather than at site is the entire point of this system-level check.

How do EPC buyers verify that a third-party inspection agency is genuinely independent?

The structure of the commercial relationship is the first signal. In rigorous EPC procurement practice, the buyer contracts and pays the TPI agency directly — the fabricator has no commercial relationship with the inspector on site. When a fabricator nominates and pays the TPI, that arrangement should raise an immediate question about independence regardless of how professional the individual inspector appears.

Beyond contract structure, EPC QA teams verify inspector credentials: ASNT Level II or III for NDE activities, CSWIP for weld inspection, and where applicable, ASME Authorized Inspector qualification for code stamps. Many EPC buyers also rotate inspectors across visits specifically to prevent the familiarity effect — an inspector who has spent six months in one shop and has a working relationship with the welding supervisor is not an independent check in any meaningful operational sense. Procurement managers reviewing TPI proposals should ask explicitly who appoints, who pays, and how inspector rotation is managed over a multi-month fabrication program.