ASME B31.3 certificate: what to send a Korean client
A Korean EPC or end user sends you a vendor document requirement list, and one line on it reads "ASME B31.3 certificate to be submitted." Your quality manager goes looking for it, finds that ASME issues nothing by that name, and comes back with a question you now have to answer in writing to a client you have not shipped to before.
Replying that no such certificate exists is accurate, and it stops the file where it is. Document review stays open, the invoice milestone behind it stays open, and spools already on the water reach a site that cannot close its receiving inspection. The request is not wrong. It is shorthand, and the way through is to send the package it stands for before anyone asks a second time.
What ASME stamps, and what it does not
ASME runs accreditation programs in which a shop is surveyed, holds a Certificate of Authorization, and applies a code symbol stamp to the product. Pressure vessels under Section VIII Division 1 get the U stamp and a Manufacturer's Data Report; power boilers under Section I get the S stamp. Power piping has the PP designator, covering boiler external piping — the jurisdictional portion of B31.1.
B31.3 has none of that. It is a design and construction code covering materials, design, fabrication, examination, inspection and testing of process piping, and it sets up no accreditation scheme, so nobody holds a B31.3 certificate from ASME. Documents with that title still circulate. Read what is printed on one rather than its header: it is either a statement the fabricator issued about its own work, or a conformity statement from an inspection body engaged to witness it. Both are legitimate to send, and neither is an accreditation.
The shorthand itself is not Korean. The same line turns up on document lists from any region. What differs here is where your answer lands: the reviewer is the client's own inspection function or an agency they appointed, the review cycle carries a payment milestone, and a question about statutory approval usually sits on the same list, a few lines below. The same records can come back for the way they are indexed, because the reviewer works down the client's own document index and submittal codes rather than the order your data book was assembled in. Whoever wrote the line means one thing: show me that this piping was built to B31.3, and give me the records that prove it.
The records that carry the proof
Compliance is demonstrated by a set of records, not a single sheet. Five do almost all the work and the sixth ties them together.
| Record | What it proves | Who signs it | What goes wrong if you substitute |
|---|---|---|---|
| Mill test certificate, EN 10204 3.1 or 3.2 | The material is the grade the line class calls for, with lot-specific results traceable by heat number | Mill QA independent of production; a 3.2 adds a second signature from the purchaser's inspection representative or a designated inspector | A 2.2 test report arrives instead. It does carry test results, but from non-specific inspection, so none of them is tied to the lot delivered, and a reviewer working to a 3.1 requirement cannot accept it |
| WPS with its supporting PQR, to ASME Section IX | The joint was welded by a procedure that was itself welded, tested and passed | The fabricator's welding authority | A WPS arrives with no PQR behind it. Section IX has no prequalified procedures, so an unsupported WPS proves nothing |
| Welder performance qualification records | The named welder was qualified in a range that covers the diameter, thickness and position actually used | The employer, who certifies and maintains the records | The qualification has lapsed, or its range stops short of the joints on your isometric |
| NDE reports with a weld map | The required examinations were performed, on identified joints, by qualified examiners | The examiner and the fabricator's QC; the owner's Inspector verifies completion | Reports without a weld map. Nobody can tie a film number to a joint on a drawing, so the whole set is unverifiable |
| Pressure test record | The system held the required pressure for the required duration under recorded conditions | The tester and the witness named on the record | A "hydro completed, OK" line in a packing note, with no medium, duration, gauge identification or calibration date |
| Certificate of compliance issued by the fabricator | The scope was designed and built to a named edition of B31.3 | The fabricator's authorised signatory | Nothing, on its own — but sent alone, without the five above, it is a claim rather than evidence |
The one place "certificate" is literally correct
Material is where the word belongs. EN 10204 defines the document types, and the distinction that matters in review is between 3.1 and 3.2. A 3.1 inspection certificate reports actual test results from the lot supplied and is validated by the manufacturer's own quality function, which must be independent of production. A 3.2 is the same document carrying a second signature beside the manufacturer's: either the purchaser's authorised inspection representative or an inspector designated by the regulations that apply. Which of the two is meant has to be named in the order — the first is not an independent third party, and assuming it is changes who you have to book and when.
Which type applies is set by the project specification and the purchase order, not by B31.3. Assuming 3.2 across a whole scope adds inspection cost and lead time to items that never needed it; assuming 3.1 on a line where the specification called for a witnessed certificate means reissuing certificates after the goods have shipped, which is harder to unwind. Read the material clause before you place mill orders, and confirm which line classes carry 3.2.
Traceability is the part reviewers actually test. The heat number has to reach the specific spool through the material transfer records and the weld map. Certificates that are correct in isolation but cannot be tied to an item number on the packing list get returned, and by then the material is in a container.
The same specification section usually governs the bought-out components — flanges, fittings, valves — where the dimensional standard rather than the piping code decides whether the assembly bolts up to what the site already has. That is its own question, and it is the one that catches ASME B16.5 flanges arriving at a Korean site.
Welding is where B31.3 hands you a different book
B31.3 does not qualify welding procedures or welders. It refers that work to ASME Section IX, which is why the welding tab of a data book is full of Section IX documents. Three things sit there: the WPS as the instruction, the PQR as the tested evidence behind it, and the performance qualification record for each welder who touched the scope.
B31.3 adds the traceability layer on top. Examination records have to identify the joint, the welder or operator who made it, and the examiner who checked it, and the employer maintains the examiner qualification records so the Inspector can call for them. In practice that is a weld register: one row per joint, cross-referenced to the isometric, the welder stamp, the NDE report and the material heat. Build it during fabrication — reconstructing it afterwards from shop travellers is a common reason a technically complete package misses the shipping date.
The examination percentage comes from the fluid service, not from the code number
"Built to B31.3" does not by itself say how much examination was required. The code sets the extent by fluid service category. Category D takes visual examination. Normal fluid service adds random radiography or ultrasonic examination of at least five percent of circumferential butt and miter groove welds, and sets its own visual extent. Category M raises that random extent to at least twenty percent, and widens what it covers — fabricated lap and branch connection welds come in alongside the circumferential butt and miter groove welds — on top of visual examination of all fabrication. Severe cyclic conditions is the one that goes to full radiographic or ultrasonic examination of them. The percentages sit in the examination paragraphs of whichever edition your specification cites, so read them there before you price the NDE.
So the useful reply to that line is a question: which fluid service category applies, and what are the design conditions on the line list? Get that designation into the order before fabrication starts, and the examination extent stops being negotiable at the wrong moment.
Inspection sits above examination. Examination is what the fabricator's examiners perform; inspection is the owner's verification that it was done, exercised through an owner's Inspector who under B31.3 may not be an employee of the fabricator or erector unless the owner is also the fabricator. On a Korean project that is normally the client's own staff or their appointed agency, and they are the reader your data book has to satisfy. Aligning the document index with what they open first is one of the things that gets settled before the goods move.
Name the edition, and check theirs
The closing document is a certificate of compliance on the fabricator's own letterhead. No ASME number, no stamp. What it must carry is the edition of B31.3 the work was performed to, because the code is revised on a cycle and a specification can cite an edition that is no longer current — or a newer one than the shop worked to. Check the client's cited edition against your engineering's, in writing, before fabrication. A difference found then closes with a documented note; found at document review, it becomes an open punch list item when nobody has time for it.
When the missing stamp is not the answer
Everything above assumes the person who wrote "B31.3 certificate" was using shorthand. Sometimes they were not, and replying that no such certificate exists is then the wrong move twice over.
The first case is scope. If your delivery includes a pressure vessel — a separator, a receiver, a jacketed drum on a skid — that item is not piping, and there a certificate genuinely does exist: a Section VIII Certificate of Authorization holder, a U stamp, a Manufacturer's Data Report. A B31.3 package covering the whole skid leaves that line permanently open. The requirement list was right; it was aimed at a part of your scope you were not reading it against.
The second case is that the line came down from a statutory route rather than from the contract design code. Korean acceptance for certain installations runs through its own approval process, and where that applies, ASME records are supporting evidence rather than the answer — no volume of B31.3 documentation discharges a requirement that was never an ASME requirement. Where that approval actually applies, and how early it has to start, decides whether this is a document problem or a schedule problem.
The tell is where the line sits in the requirement list. Under a QA/QC document index, with submittal codes and a review cycle, it is the code package described here. Under a permitting, authority or local-approval heading, addressed to a party that is not the client's own inspection function, it is something else, and the answer starts months earlier. Send us the line list and the fluid service designation from the enquiry, and we will come back with the document index a Korean reviewer opens first.