Can We Perform UT At 50% Completed Weld?

[Aliq298]

Dear All Expert,

I got 1 query regarding UT Scan at 50% completed weld? Purpose i want to conduct UT at 50% completed joint due to it is critical joint and want to avoid defect on early stage. The base material thickness is 50mm, weld completed around 20mm.

Newbie,

Regards,

 

[pradipgoswami]

Hello Salleh86,

If the technique is conventional pulse echo technique then there would be little success. However with advenced techniques such as TOFD or Phased Array, inspection of partly filled weld (groove weld,I presume) would be possible. Advice would be to calibate the equipments on the similar joints before actual inspection.

Thanks.
Pradip Goswami,P.Eng.IWE
Welding & Metallurgical Specialist
Ontario, Canada.
Email-pgoswami@sympatico.ca,
pgoswami@quickclic.net

 

[racookpe1978]

More of the problem will be - as noted immediately above - the "calibration" of the partial joint "in the mind of the reader" who is looking at the ultrasonic test.

See, assume you have a full penetration "good" weld and pass a UT signal into that "good" weld joint. The reader - the person who actually is going to accept or reject the joint based on the signal that is NOT reflected from the "perfect" weld - is going to know what to expect, right? So, if one the second joint, there is a flaw, the reader is going to recognize the flaw - BECAUSE he or she sees he difference between a good joint and a flawed joint - and is going to respond as you expect. He (or she) is going to reject the "bad" joint reliably and is going to accept the good joints reliably, right?

Now, you are going to have an irregular, partially consumed, partially "open" unmelted joint that is going to be reflecting "weird" and "different" signals from what might be a flaw, and from what might be just an incomplete part of the weld where the two pieces of metal are not yet melted together. Remember, you are "planning" to be UT testing a 50% joint, but part of that joint is really going to be part 37%, 45%, part 50% and part 55%, and maybe even as much as 75%, so all of your weld is going to differ in different areas.

You might be seeing flaws, and have a good chance of cutting them out before you continue. But you might miss flaws. And you might be cutting out good weld as well.

If you are very, very concerned with inside weld quality, and have the budget, get a 50% X-ray. It will be more clear.

"Calibration" would in this sense, be a matter of taking identical weld geometry joints and doing a series of 10 50% welds, half with deliberate flaws and half with "perfect" welds. Then, UT testing each of the ten - twice with two different readers and two different UT specialists! - and see if you can tell the difference in the 20 readings between good and flawed welds.

 

[stanweld]

We have used phased array with great success in applications similar to yours. It is has proven preferable to RT.

 

[Duwe6]

"We have used phased array with great success in applications similar to yours. It is has proven preferable to RT."

Concur.

Also, with an experienced UT tech that knows welds and weld flaws [not just looking for reflectors bigger than Reference calibration], manual UT shearwave will be about as accurate as PA at 50%. With that much weld done - 20mm / .7" - the tech will get a good look at the root. YOU will prefer PA, because the scanning is easily displayed on a screen, in a 2-dimensional representation and color-coded as to percent of Reference. Fairly easy for you to 'read' and see what the tech is seeing.

With manual UT, when the tech says 'good' or 'fix it', you will just have to trust him. If you keep the same tech long enough, and he/she finds a few 'fix it's' [few and far between with good welders], you will be able to see the defects the tech rejected, if you watch the repair process.

 

[phighfill]

If you have a "good" shearwave hand that understands what is going on then that should suffice. I have had many years at doing this. It all boils down to the calibration and EXPERIENCE.

 

[ndeguy]

What's the form of the component - unlikely to be pipe at 50mm wall thickness. So if there is access to the root face and you have let the joint cool down, then visual inspect to check for lack of penetration/root condition followed by MPI for any lack of fusion. follow this by manual UT from the root face surfaces (both sides of the root). You can use a good 70 degree probe and it should give clear indications of any lack of fusion defects down to the free weld surface at 20mm.