Vendor Surveillance for Process Equipment and Piping Fabricated in the Bangkok Region
Third party inspection in Bangkok places a dedicated inspector inside your supplier's shop against a purchase order: agreeing the ITP, witnessing pressure-part hold points, reviewing radiography and heat treatment charts, attending hydrotest, and releasing only against a complete manufacturing record book. Atlantis inspectors mobilise from Houston or Hyderabad and remain deployed for the whole contract.
Thai fabrication capacity for process equipment is concentrated in a corridor that runs south-east from Bangkok through Samut Prakan and Chonburi to the Rayong and Map Ta Phut petrochemical complex, with the port at Laem Chabang at the end of it. The shops there build pressure vessels, columns, heat exchangers, storage tanks, piping spools and packaged skids, many of them holding ASME certificates of authorisation, and much of the output leaves on a project schedule that has no slack in it. The commercial risk for a buyer is rarely gross workmanship. It is the accumulation of small, undocumented departures: a substituted heat, an unrecorded repair, a heat treatment chart with a gap in it, a hydrotest attended by nobody. Surveillance exists to convert those into signed evidence at the moment they happen, when correction still costs shop hours rather than site downtime.
Source: ASME BPVC Section VIII Division 1, including UG-77 marking, UG-93 inspection of materials, UW-33 alignment tolerances, UW-51 radiographic examination, UCS-56 postweld heat treatment, UG-84 and UCS-66 impact test requirements, UG-99 hydrostatic test, UG-116 and UG-119 marking and nameplates; ASME BPVC Section IX welding qualification; ASME BPVC Section V Articles 2, 6, 7 and 23; ASME B31.3 Process Piping, including para 323 materials, Table 330.1.1 preheat, para 331 and Table 331.1.1 heat treatment, para 344 examination and para 345 pressure testing; SSPC-PA 2 for dry film thickness measurement; ISO 8501-1 and ISO 12944 for surface preparation and protective coatings.
| Fabrication stage | Point type | Evidence the point must produce | Governing reference |
|---|---|---|---|
| Material identification and traceability | Hold | Heat and lot numbers transferred to every cut part before the original mark is lost, mill certificates matched to the material list, positive material identification on all alloy pressure parts | ASME VIII Div 1 UG-77 and UG-93; ASME B31.3 para 323 |
| Welding procedure and welder qualification | Hold | WPS and supporting PQR checked against the actual joint configuration and thickness, essential variables verified, welder continuity within period, unique welder identity assigned on the weld map | ASME Section IX, QW-200 and QW-300 series |
| Fit-up, preheat and interpass | Witness | Root gap and permitted misalignment, joint cleanliness, preheat measured by contact pyrometer at the specified distance from the joint, interpass temperature controlled and recorded | ASME VIII Div 1 UW-33; ASME B31.3 Table 330.1.1 |
| Nondestructive examination | Witness or review | Technician certification and current vision record, equipment calibration, radiographic density and IQI sensitivity, technique sheet matched to the joint, repair mapping and full re-examination after every repair | ASME Section V Articles 2, 6, 7 and 23; ASME VIII Div 1 UW-51; B31.3 para 344 |
| Postweld heat treatment | Hold | Continuous chart trace with no gaps, thermocouple count and attachment locations, heating and cooling rates through the controlled range, soak temperature and hold time against governing thickness, chart identified to the item | ASME VIII Div 1 UCS-56; ASME B31.3 para 331 and Table 331.1.1 |
| Impact testing and MDMT | Review and hold | Minimum design metal temperature on the data report reconciled against actual service, exemption curve applied to the correct material group, impact test results where exemption does not apply | ASME VIII Div 1 UG-84 and UCS-66 |
| Hydrostatic or pneumatic test | Hold | Calibrated gauge with valid certificate and correct range, test medium and chloride content for austenitic stainless, hold duration, examination at pressure, vents and blinds verified, immediate draining and drying | ASME VIII Div 1 UG-99; ASME B31.3 para 345 |
| Nameplate and stamping | Hold | Nameplate data reconciled line by line against the data report, stamping applied only after the Authorized Inspector has accepted the vessel, registration arranged before shipment where the destination jurisdiction requires it | ASME VIII Div 1 UG-116 and UG-119 |
| Surface preparation and coating | Witness | Blast profile and soluble salt testing, steel temperature and dew point margin at application, dry film thickness against specified minimum and maximum, adhesion and holiday detection | SSPC-PA 2, ISO 8501-1, ISO 12944 and the project coating specification |
| Final documentation and release | Hold | Manufacturing record book indexed and complete, all non-conformance reports closed with verified disposition, release note issued against the purchase order line item | Purchase order, project quality specification and the U-1A data report |
Placing inspection against a purchase order at a Thai fabrication shop
Vendor surveillance is bought the way the equipment is bought: against a purchase order, with a defined scope, a defined visit pattern and a defined deliverable. For process equipment that usually means a monthly or campaign-based engagement covering several order lines at one or more shops, running from material receipt through to release, rather than a single final inspection visit at the end.
The scope divides into four activities that should be priced and reported separately. Document review covers the ITP, drawings, procedures and qualification records before production. Witness and hold point attendance covers the scheduled events. In-process surveillance covers unannounced presence between them. Expediting covers schedule and reports on delivery risk. Confusing the last with the first three is the most common structural error in a surveillance contract, because it produces reports that satisfy neither purpose.
What the buyer gets from this is evidence rather than reassurance. Each intervention produces a signed record naming the requirement, the observed condition and the disposition. That record set is what makes a claim defensible, an audit survivable and a warranty argument short.
The pre-inspection meeting: aligning specification, drawing and shop reality
Hold the pre-inspection meeting at the shop, before production release, with the shop's welding engineer and QC manager present rather than only its commercial team. The objective is not to read the specification aloud. It is to find the places where the shop's understanding and the contract diverge while the divergence is still free to correct.
The specific items to resolve are the ones that later become disputes. Which drawing revision is on the shop floor and does it match the one issued to you. Which welding procedures cover the actual joints, particularly branch connections, nozzle-to-shell welds and any dissimilar metal joints. Whether NDE and heat treatment are performed in-house or subcontracted, and whether the subcontractor is approved. What the notification route is, in writing, with named individuals. What constitutes readiness for a hold point, so that an inspector does not travel to find the joint unprepared.
The other output is the ITP with point status agreed line by line. Assigning hold status generically to everything is as unhelpful as assigning it to nothing; it guarantees waivers when the schedule tightens. Assign hold status where the evidence is destroyed by the next operation and witness status elsewhere. Where the underlying examination procedures themselves need writing or challenging, ASNT Level III consulting provides the technical authority to do that independently of the shop performing the work.
Codes that govern process equipment and piping made in Thailand
Most process equipment fabricated in the Bangkok and Rayong corridor is built to ASME Section VIII Division 1 with welding qualified under Section IX and examination performed under Section V. Piping is normally ASME B31.3. A number of Thai shops hold ASME certificates of authorisation and can apply the U stamp, and that certification is genuine and worth having, but it certifies the shop's quality system rather than any particular vessel.
Two code items are underspecified by buyers with predictable consequences. The first is the minimum design metal temperature. Buyers frequently accept whatever MDMT the fabricator's design produces without reconciling it against actual service conditions, including upset and depressurisation cases. The nameplate MDMT then turns out, at commissioning, to be warmer than a credible operating case, and the options at that stage are all bad. Settle MDMT and the impact testing consequences at design review, not at nameplate.
The second is the examination percentage on piping. B31.3 normal fluid service requires a defined proportion of girth welds to be radiographed, distributed across welders and procedures, and severe cyclic or Category M service imposes considerably more. Shops sometimes satisfy the percentage by shooting the accessible joints repeatedly rather than sampling across welders. Ask to see the distribution, not just the count.
Alloy piping: PMI, heat traceability and the Grade 91 trap
Alloy piping is where material control failures concentrate, because P11, P22, P91 and the austenitic grades look identical in a rack and are frequently stored together. Positive material identification should be one hundred per cent on alloy pressure parts, including fittings, flanges and weld consumables, not a spot sample. A single carbon steel elbow in a chrome-moly line is a failure that will not appear until the plant is hot.
Grade 91 deserves its own attention because it is unforgiving of heat treatment error in both directions. Its post-weld heat treatment window is narrow, and the consequences of missing it are opposite at each end: soak below the intended range leaves untempered martensite and hardness above the specified limit, while excursion above the lower transformation temperature re-austenitises the material and produces a joint that cannot be corrected by further tempering. Both look identical on a completed weld and both are found by hardness testing, which is why the hardness survey is not optional paperwork.
The associated document control failure is furnace charts that cannot be tied to a specific item. A chart without the item number, the furnace identity and the thermocouple layout is not evidence, however good the trace looks. Insist that heat treatment records are item-traceable at the point they are produced, because reconstructing that link at data book stage is guesswork.
NDE review: the part buyers most often accept on trust
Radiography is where the largest volume of examination evidence is produced and where the least buyer scrutiny is applied. A meaningful review checks the technician's certification and current vision record, the equipment and source calibration, the technique sheet against the actual joint, the image quality indicator and its essential hole or wire visibility, the density within the acceptable range for the source used, and the identification and location markers that tie the image to a specific joint on a specific item.
The finding that matters most is repair traceability. Every repair must be radiographed again over the full repaired length plus the required overlap, and the repair sequence must be traceable through the weld map. Where a repair rate climbs above a few per cent, that is a process signal rather than a workmanship one: a consumable problem, a fit-up problem or a welder outside their comfortable range. Treating it as a paperwork issue misses the opportunity to fix the cause while the order is still in the shop.
Reviewing a large volume of radiographic and other NDE reports properly is specialist work and is normally worth separating from the surveillance visit itself. Atlantis provides independent review of inspection reports as a distinct service, which gives a buyer a second technical opinion on the examination record without slowing the visit schedule.
Hydrotest, nameplate and the moments that cannot be re-run
The hydrostatic test is the single most consequential witness point on a vessel because it cannot be reconstructed afterwards. Verify the gauge before pressurising, not after: correct range so that test pressure falls in the useful portion of the scale, valid calibration certificate, and a second gauge where the specification requires it. Confirm the test medium, and on austenitic stainless confirm the chloride content of the water against the project limit, together with the requirement to drain and dry immediately after test.
Then watch the test rather than the paperwork. Hold duration at pressure, examination performed at the reduced examination pressure rather than at peak, vents open and confirmed, temporary blinds rated for the pressure, and the pressure trace recorded. A hydrotest certificate signed by nobody who was present is a document, not evidence, and that distinction becomes very sharp during a failure investigation.
Nameplate verification closes the sequence. Reconcile every field on the nameplate against the data report line by line, confirm that stamping followed acceptance by the Authorized Inspector rather than preceded it, and confirm that registration with the destination jurisdiction has been arranged before shipment where it is required. Retrofitting a registration after a vessel has left the shop is expensive and sometimes impossible.
Expediting against delivery and the shipping window
Delivery risk on Thai process equipment orders concentrates in three places, and none of them is welding. The first is long-lead material: plate in the required thickness and grade, forgings, bolting and specialty valves, all frequently sub-ordered late and all capable of stopping a shop that is otherwise ready. The second is subcontracted operations, principally heat treatment and NDE, where the queue belongs to somebody else's schedule. The third is the shipping window itself, since a vessel that misses a booking at Laem Chabang waits for the next one.
A useful expediting report states, for each order line, the current fabrication stage against the plan, the status of every open sub-order with a promised date, the manpower assigned, the current repair rate, and a stated view on the delivery date with the reasons. It does not simply repeat the shop's forecast. The value in having the same inspector on site each week is that the forecast can be tested against what the bay actually looks like.
The regional schedule fact worth planning around is Songkran in mid-April, which closes shops for several days and reduces output either side of it. It is entirely predictable and routinely absent from delivery schedules written elsewhere. Raise it at order placement rather than expediting against a date that was never achievable.
The manufacturing record book and what release actually certifies
The record book is the deliverable that outlives the inspection campaign by thirty years. It is what an owner's integrity engineer will open when planning the first internal inspection, what a jurisdiction will ask for at registration, and what an insurer will ask for after an incident. Build it progressively through fabrication and check it against the contract document list at defined stages, because assembling it in the final week guarantees gaps that can no longer be filled.
Release is a documentary act with a narrow meaning. The inspector signs a release note when the ITP is closed point by point, all non-conformances are closed with verified disposition, and the record book is complete and indexed. The release note does not certify fitness for purpose, and it should never be drafted as though it does. That distinction protects the buyer as much as the inspector.
Where a buyer operates a fleet of equipment rather than a single purchase, the record book is the seed of the asset's inspection history and should land somewhere structured rather than in a folder of scanned PDFs. The design data, material certificates and baseline thickness readings captured at fabrication are the reference every future inspection is measured against, which is why they belong in an inspection data management system from day one.
From fabrication release into the owner's integrity programme
The most valuable output of a well-run surveillance campaign is not the release note. It is the baseline. Nominal and actual thicknesses, weld maps, material certificates, heat treatment condition, design MDMT and corrosion allowance are all determined at fabrication and all needed at every subsequent inspection interval for the life of the equipment. When they are lost between the fabricator and the operating site, the first turnaround inspection has to reconstruct them at considerable cost and reduced confidence.
This is the strongest practical argument for holding surveillance records in a structured form rather than a document archive. The equipment number, the design conditions and the as-built thickness survey feed directly into the corrosion rate calculations and remaining life estimates that drive the owner's inspection planning under its mechanical integrity programme.
Continuity of personnel helps here too. An inspector who has been deployed on the campaign since material receipt knows which nozzle was re-welded, which chart covers which item and which non-conformance was closed by analysis rather than repair. That knowledge belongs in the handover, and it survives only if it was written down at the time by someone who was actually present.
How Atlantis staffs a Bangkok surveillance campaign
Inspection teams mobilise from Houston or Hyderabad to the supplier's works and remain deployed for the duration of the contract. One inspector, present across the campaign, who knows the specification, the drawing revisions, the shop's welding engineer and the open non-conformances. That continuity is the difference between a surveillance programme and a series of unrelated visits, and it shows up commercially in the second month, when the interpretation arguments have already been settled and recorded.
Coverage flexes with the work. A campaign can intensify around material receipt, heat treatment and hydrotest and reduce during periods of low shop activity, without changing personnel or re-explaining the specification. Reports are issued the same week as the visit, non-conformances are raised on the day they are found, and the record book is built progressively rather than assembled at the end.
Inspectors are ASNT and ISO 9712 qualified in the relevant methods and work to the client's specification. To scope a campaign against specific Thai suppliers, a consultation is the starting point, and we work back from your purchase order lines, the codes they invoke and the delivery dates you have committed to.
What does a pre-inspection meeting achieve before fabrication starts?
It reconciles three documents that are rarely identical: your purchase order specification, the drawings as issued for construction, and the shop's own quality plan. In one session you fix the ITP point status, notification route and notice period, confirm which welding procedures cover the actual joints, agree the NDE technique sheets, and settle how non-conformances will be raised and dispositioned. Every argument avoided here is a delay avoided at the hydrotest.
Why does a post-weld heat treatment chart need reviewing line by line?
Because the chart is the only evidence the cycle happened as specified, and it fails in quiet ways. Check the continuous trace for gaps, the thermocouple count and attachment locations, heating and cooling rates through the controlled range, soak temperature against the material, and hold time against the governing thickness. Confirm the chart carries the item number and the furnace identity. A chart with a flat, suspiciously perfect soak is worth a harder look.
What should a manufacturing record book contain for an ASME vessel?
An indexed set covering the U-1A data report, material certificates matched to marked heat numbers, PMI results on alloy parts, welding procedure and performance qualifications, the weld map with welder identity per joint, consumable certificates, NDE reports and radiographs or digital images with the repair history, heat treatment charts, hardness results, hydrotest record with gauge calibration, dimensional report, nameplate rubbing, and every non-conformance report closed.
How does expediting differ from inspection on a Thai fabrication order?
Inspection asks whether the work conforms. Expediting asks whether it will arrive. The second requires visibility of sub-orders for plate, forgings, bolting and valves, of shop bay loading against competing orders, of NDE and heat treatment subcontractor queues, and of the shipping window at Laem Chabang. A single visit that does both is efficient, but the reports must stay separate, because a schedule report signed as an inspection report confuses the record.
Is third party surveillance the same as the ASME Authorized Inspector?
No. The Authorized Inspector is a statutory role held by an Authorized Inspection Agency, and only that inspector accepts the vessel for Code stamping and signs the data report. Third party surveillance is the purchaser's own verification, carried out against the purchase order and the project specification, and it covers a great deal the Code does not: coating, dimensional interfaces, preservation, documentation completeness and schedule. The two run alongside each other.
What goes wrong when a hold point is waived to protect the delivery date?
The waiver moves the discovery, it does not remove the defect. A missed fit-up point becomes a misaligned nozzle found at site. An unwitnessed hydrotest becomes a dispute over whether the gauge was calibrated. An unattended PWHT becomes a hardness result nobody can explain. Each is recoverable in the shop for hours of labour and unrecoverable at site for weeks, and the waiver almost always originates with the buyer, not the shop.