Repair Welding on a Joint Already Accepted
Radiography of a completed circumferential butt weld in a 25 mm carbon steel header found three indications: two rounded indications assessed as porosity and one planar indication assessed as lack of fusion on the root. The joint had already been accepted on visual and surface examination and signed off by the supervisor. The client's engineer wants the whole weld cut out and replaced. The fabricator wants to grind out the two porosity areas and leave the lack of fusion, arguing it is 'root-side and inaccessible'. Repair would be by the same procedure, in the same position, by the same welder. The header is on the critical path; replacement adds eleven days.
1. How should the supervisor assess the two types of indication differently, and what does each require?
Read the worked answer
The two are not equivalent. Rounded indications assessed as porosity are volumetric, generally distributed, and usually tolerable within acceptance limits or removable by local excavation. The planar indication assessed as lack of fusion is a crack-like planar defect on the root; it concentrates stress, does not blunt under load, and is the defect class most associated with premature failure in service. It cannot be left.
The supervisor should also address the 'inaccessible' argument directly: lack of fusion at the root of a 25 mm butt weld is not inaccessible to excavation from the root side if the joint is open, and if it is not, that is a constraint on the repair method, not a reason to accept the defect. Acceptance is decided against the specified acceptance criteria, not against what is convenient to remove.
The correct position: excavate and repair the lack of fusion, and verify by examination after repair. Deal with the porosity on its merits against the acceptance criteria; it may not require repair at all if within limits.
2. Repair by the same welder, same procedure, same position: is that acceptable, and what must the supervisor verify before authorising it?
Read the worked answer
It can be acceptable, but the supervisor must not authorise it automatically. The first question is why the defect occurred. Lack of fusion at the root typically points to insufficient heat input, poor joint fit-up or gap, wrong technique at the root pass, or inadequate cleaning. If the cause is the procedure or the welder's technique, repeating both reproduces the defect.
Before authorising repair the supervisor should: confirm the excavation is complete by examination (the defect must be gone before welding, not buried); check the welder's qualification covers repair welding on this material and thickness; confirm the WPS is suitable and, if the cause was procedural, revise it; ensure preheat and interpass requirements are met for 25 mm carbon steel; and plan post-repair examination at the same level as the original.
A welder who produced a planar root defect should be assessed, not simply re-tasked. Repeating the same welder on the same procedure without investigating cause is how a repair becomes a second defect.
3. The client wants full replacement, the fabricator wants minimal intervention. How should the supervisor justify a position to both?
Read the worked answer
By separating the technical question from the commercial one, and answering the technical question on evidence.
The technical position: the lack of fusion must be removed and the repair verified; the porosity must be assessed against the acceptance criteria and repaired only if it exceeds them. Full replacement is not technically required if a verified repair restores the joint to the specified condition. That is the answer to the client, supported by the acceptance criteria and the proposed examination.
The commercial position is not the supervisor's to trade. Eleven days of delay is a real cost, but it does not change what the joint requires. The supervisor should state the technical minimum clearly, offer the verification evidence that will demonstrate the repair is sound, and let the parties decide about schedule.
What the supervisor must not do is sign off a joint containing known lack of fusion because the schedule is tight. That signature would be false, and it would be indefensible if the header failed.