ASME B31.3 Para. 345.2.3(c) Closure Welds

[AEF]

According to B31.3 Para. 345.2.3(c) for closure welds, this may be done as long as all components have been leak tested successfully, undergo in-process examination as per Para. 344.7, and pass either 100% RT or 100% UT examination. This becomes a problem when the final closure weld is a socket-weld. Has anyone been able to develope and qualify either a UT or RT procedure that will satisfy the requirements for closure welds when they're a socket weld configuration? We've submitted test pieces, consisting of 1" sch. 80 pipe into 3000# SW ell's, to a few NDE companies, with level 3's, asking if they can develope a procedure for either RT or UT and so far none have been able to do so. Anyone else have any better luck?

 

[Seldom]

Hi AEF,
I’ve been in your situation several times with socket closure welds and have taken this stance:
I performed the in-process examination and used MT on each layer of weld metal inleu of using either UT or RT.

As you’ve discovered the limitations of UT and RT pertaining to socket welds, I suggest to you that while the requirements of 345.2.3(c) are basic design principles, the Code is not a design handbook by it’s own admission. I further suggest that the use of MT as I’ve stated in this scenario, constitutes competent engineering judgment and assures uniform application of the Code’s safety and quality principles.

I believe now as then, this is a commonsensical approach to this specific problem and would have no problem enacting it again under like circumstances.
Good luck to you.

 

[Seldom]

Oops, I forgot to mention in the previous post (clicking too quick) an additional reason to use MT or PT. Please take a look at the acceptance criteria table 341.3.2 and check out the fillet weld imperfections listed. Notice which ones are NA's and which ones are detectable using MT or PT.

 

[moseley]

I have been in the same situation many times and have found that with client acceptance (which has never been a problem) I have performed PT on each pass. One at the "root" and one on the final pass. We have settled many issues that way. I have found that this is the best procedure to satisfy the code. RT and MT supply minimal results on this type of joint. Hope this helps.