If a weld on a WFI loop cannot be traced to a borescope image and a coupon, nobody can defend it at IQ.
Tying one weld to its evidence takes five records, all keyed to the same weld number:
- A weld map on the isometric, with every weld numbered once and never renumbered.
- Welder and machine identity against each weld number, so a problem with one operator or one power source can be traced to every joint they touched.
- The parameter printout from the orbital power source for each weld, filed under its number.
- Coupons made at the start of each shift and after a setup change, examined for full penetration and discolouration before production welds continue.
- Borescope images named by weld number. An image called IMG_0453 proves nothing.
ASME BPE sets the acceptance criteria for hygienic tubing welds, covering penetration, concavity, discolouration and more. Those criteria only help if the evidence can be matched to the weld it describes.
When the image and the map disagree at IQ, the options are poor. If the line is still open, the weld can be scoped again. If it is insulated and supported, the usual answer is to cut the weld out and make it again, with a new coupon, a new printout and a new image, on a system that was otherwise ready.
On the equipment we supply, weld and passivation records go into the equipment file while the work is being done, so IQ starts from a complete record.
Topics


