A heat number identifies a batch of metal produced together, and it is the only reliable link between a finished part and the mill certificate that describes the metal it was made from. It does not identify a grade in general. It identifies one specific melt, with its own chemistry and its own history of processing.
That distinction matters because two heats of the same grade can behave differently. One may form and weld cleanly, the other may crack at a formed radius that has never been a problem before. Without the heat number, a shop has no way to connect a sudden run of failures to a change in incoming material.
The chain breaks at the saw. Material arrives with the number on a tag or a stencil, gets cut to length, and the offcut goes into a rack with forty other offcuts. From that point, the number on the tag refers to a piece that no longer exists in that form, and the identity of the cut piece is whatever someone remembers.
Every step after cutting multiplies the ambiguity. Forming, welding, finishing and assembly all move the part further from its original marking. By the time it reaches dispatch, the number has either travelled with the part in some durable form or it has been lost, and there is no reconstructing it afterwards.
The moment of cutting is the moment to act. A tag transferred to the offcut, a mark on the end, a bin label that stays with the batch: any of these keeps the chain intact. Doing it a shift later is guesswork dressed as paperwork.
The awkward cases are the small ones. A short offcut of titanium tube has nowhere convenient to hold a tag, and it is exactly the piece that gets used for a bracket nobody is tracking. Deciding in advance how small pieces will be identified is part of the process, not an afterthought.
Marking has to survive the process the part will go through. Ink disappears in a solvent wash. A sticker comes off in a degrease tank. A hard stamp in the wrong place creates a stress riser, which is a particular concern on thin-wall titanium and on anything that will see vibration.
The practical answer is usually a combination: a low-stress mark in a non-critical area, transferred to a bag or a bin at each stage, plus a record on paper that ties the batch to the job. None of those alone is enough, and together they cover the ways identity is normally lost.
A mill certificate that sits in a folder proves nothing about any particular part. The proof is the connection: this certificate, this heat, this cut length, this welded assembly. Each link has to exist in a document that someone can follow.
The simple version is a job sheet that lists the heat numbers of the material issued to it, alongside the quantity and the part reference. When a customer asks what the part is made from, the answer is a chain rather than an assurance, and the chain can be checked.
Batch traceability says every part in this group came from this material. Serial traceability says this specific part did. The first is enough for most fabricated work; the second is needed where a single component carries a safety consequence and a failure would have to be investigated individually.
Choosing between them is a commercial decision as much as a technical one, and it should be made at the quotation stage. Building serial traceability into a batch job is expensive; discovering halfway through that it was needed is worse, because the earlier parts cannot be reconstructed retroactively.
The most common real-world failure is not a lost certificate. It is two heats of visually identical material on the same bench at the same time. Once the pieces are cut and the tags are off, the only thing separating them is the memory of whoever picked them up.
Physical separation is the control that works. One heat per bench, or clearly divided storage with the number visible from the working position. Rules that depend on remembering which pile is which will eventually fail, and they fail silently, which is the worst way for a control to fail.
A traceability record should be short enough to complete during the job and complete enough that a stranger can follow it. The test is whether someone could answer a customer's question in a year's time without interviewing anyone.
Reconciliation is the part most shops skip. If more material was issued than the parts account for, the difference is either scrap or a part with unknown parentage, and knowing which is the point of the exercise.
Traceability gets a reputation for paperwork because it is usually implemented as paperwork. The version that works is built into the flow: the number is on the tag, the tag goes with the cut, the job sheet has a field for it, and the field is mandatory because the job cannot close without it.
Done that way it costs a few seconds per batch and it answers questions that would otherwise be unanswerable. It also changes behaviour, because a welder who knows the material's origin is recorded treats an unfamiliar heat with a little more attention.
The same approach carries across product families. On additional fabricated parts the batch record is usually enough, while titanium welding parts often need the material and the procedure tied together in the same file.
For welded aluminium tanks the traceability question usually becomes a safety question, and for fabricated titanium components it becomes a corrosion question, because contamination and parentage are linked. Our custom work and company profile pages describe how that is handled in practice.
A buyer who asks for a sample traceability pack before placing an order learns two things at once: whether the shop can produce one, and whether producing it took an hour or a minute. The second answer is the more informative. Ask for the pack from a recent job rather than a showpiece, and check that the heat number on the certificate matches the one on the job sheet.
