NDE of a set-in nozzle weld in ASME VIII div 1 thin duplex shell 2

[mg75]

Hi everyone,

After reading through ASME VIII div 1, some of its code cases, ASME V, and various papers about the subject, I am trying to make up my mind about the most convenient way to inspect the welds connecting set-in nozzles to their duplex vessel.

I would usually stick to the code requirements, however since some welders were just qualified and have little experience in welding duplex, I am nervous about the actual quality of the welds. Some defects also detected by RT in shell butt welds did not help to alleviate my concerns. So I would like to perform some NDE on these nozzle welds, but I could not find any which would be completely satisfactory.

First, the vessel is 8mm thick, in duplex. The nozzle, made either of duplex tube or of duplex rolled sheet, would be roughly 5mm thick. Assembly is as per ASME VIII div 1 Fig UW-16.1 (c) or (d).

According to table UW-12, RT are not applicable. Does it mean it is just not required, or is it also because such type of joint does not permit interpretable radiographs ?

Sometimes, US are considers as a alternative to RT. However I am told that the thicknesses here are too thin, using either phased-array or straight beam (from inside the nozzle).

Does anyone have a suggestion about a test that may give a reliable picture of what the whole weld looks like (i.e. not only dye penetrant)?

Thanks for reading and for your advice.

 

[metengr]

First, RT is not required for this application.

I would try digital RT.

 

[Vinch]

Ultrasound testing of welds in 8 mm thick material is feasible. However, duplex steel grain structure of weld HAZ is more of a challenge for UT. Maybe a highly qualified technician could come up with a reliable test procedure.

 

[MikeMet]

We have had luck using phased array UT of nozzle fillet welds in duplex vessels. You will need a proper calibration sample to ensure relevant results.

 

[r6155]

1) You must think in design by inspection. Next time think with UW 16 (f-1) or (f-4)

2) Penetrant test should be used in each welding filling secuence, first to last pass.

3) Try with Acoustic Emission Examination

My opinion only

Regards

r6155

 

[mg75]

Thank you all for your replies.

I just tell the end of the story, for those who would read this topic later on.
We decided to perform a destructive examination on a nozzle, and found out that nozzle fit-up was probably different from what was designed.

I think we saved time compared to the UT option, because it has to be calibrated on a sample close enough to the actual fit-up, with calibrated flaws. If the calibration block and the actual arrangement are too different, it will be impossible to get any reliable information from this examination.

In the end, we repaired all the nozzles already mounted and noted that the fits-up were more or less questionable.

Next time, we plan to monitor thoroughly the first nozzle fit-ups and welds to prevent this, with the manufacturers we trust the less, in addition to the usual PT.

My conclusion : do not fear destructive examination if you fell uneasy with the NDE.

 

[Duwe6]

Set-in nozz's are easily inspected with UT. vinch & MikeMet both have very valid points, though

Accurate calibration is necessary for the duples weldmetal. One of the UT cal standards has to have a reference notch in, or on the other side of, a substantial weld.

Manual UT is perfectly adequate, if the UT tech is experienced. Phased Array provides color-coded results that are much easier to interpret, and allows you easily see what the PA tech is finding. Just look over his/her shoulder, and PA gives you all the results pictorially.

 

[EuroWeld]

Why so affraid of made welds? Welders are qualified right? Give than some time to practice! Why don't you perform penetrant testing after each welded layer and final surface? Much cheaper!
Is welding duplex so much different than welding normal austenitic stainless steel?

I'm curious,

 

[Duwe6]

When duplex is welded 'cold' enough to avoid cracking, the welder may deposit a bead that is so 'cold' that it fails to fuse with the sidewall of the basemetal; Lack-of-Fusion. UT detects it well, RT has a very poor probability of detection. PT cannot detect 'tight' L-o-F, as it is a braze joint.

100% Fit-up inspections, and regular & random in-process welding inspx are the norm in a GOOD vessel shop. Cheaper than destructive testing, usually cheaper than 100% UT.