Repads for Pressure vessel

You obviously have a leak path for air to reach between the two repad halves in the space formed between the repad and nozzle surface. The fillet weld surrounding the repad halves is OK from what you indicated above during your pressure test.

However, you did not explain or mention how you joined the two repad halves along the split line. My guess is this is not a full penetration weld and this is where the air is going between the two repad halves when you pressure test.

You likely have full penetration attaching the 2 plates but lack of fusion attaching to the shell. I don't see this as being a problem providing you test each side with the other plugged off.

codeeng,
Sounds like you think that the weld defect is acceptable that allows leakage to the other side of the repad. I'd be concerned that the lack of penetration of the repad butt weld into the vessel shell could be a crack-like defect. I suggest that it be investigated as to nature and extent.

Steve Braune
Tank Industry Consultants

SteveBraune
I'm no more concerned than if this were a single plate with the void all round. Probably do more damage removing the weld(s).

codeeng,
Nor would I if that were true. The leak is a sign of a weld defect of which its character can not be known. Since many weld defects end up crack-like in nature, the leak should not be accepted without further verification.

Steve Braune
Tank Industry Consultants

I would not call this a weld defect, since there is no obligation to weld to the shell at these joints. There are many other welded joints subject to zero inspection.

I agree with Steve. A leak shows a defect and it was not expected according to the first post. Now that you know it is there I see it as negligent not to find out why.

Could it be a material defect/lamination? Maybe it is just a weld defect,but if this was designed as a full pen weld this is not acceptable.

To me it is a different matter about acceptable levels of inspection. In my book this would require a nonconformance and a evaluation.

I also have to agree with deanc and SteveB on this matter.

I must agree with codeeng: what is required to be full pen is of course only the weld between the two halves of the repad. For this weld vessel wall acts only as a back support: its role, together with a correct bevel geometry and gap, is only to allow a correct filling of the joint in order to obtain the full penetration.
If it was possible to avoid that this weld fuses also part of vessel wall, I think this would desirable, to avoid possible damage to the main wall. A copper backing strip is an example, though this would create a highly undesirable space between the repad and the wall: however it is clear that in such case air would freely traverse from one hole to the other without any concern.
There are so many problems in building a good vessel: no need to create new ones.

prex

Online tools for structural design

prex: Your point is valid. Perhaps I did not explain myself well enough.

From the first post it appeared as they did not expect a leak. So...I ask. Did the design call for the repad joining weld to be attached to the shell? If so then there may be another defect. The repad may be fine by itself as you suggest. I see another issue.

If I had to make a call on this I would return to the weld joint design for what was specified. If it called for,say a V grove with a 1/4" open root. I would call the leak a problem.

For a full penetration between the repad halves, assuming that the shell was used as a backing bar, should result in a no leak joint when pressure testing to repad. If that is what was intended, then a leak suggests a weld defect is present. Said weld defect may be simply a pinhole, a weld crack, a crack-like defect or a lack of penetration. Until that is known, you don't know whether its acceptable or not.

Steve Braune
Tank Industry Consultants