UT of Carbon to Stainless Weld - Heat Exchanger

[compit]

Hi,

We have the final closing weld on a closed heat exchanger to ultrasonically test. (There is no access for radiography.)

We can not find any UT procedure to test this weld.

There are obvious complications with the change in speed and interface boundary between the two materials so special considerations will have to be made.

Would anyone have any advice on an approach to take?

Thanks
Tom

 

[Duwe6]

To do this 'right', you would need a carbon-to-stainless test plate with the same, or thicker, thickness. Have artificial flaws -- notches -- EDM'd into the ID and OD, in both [c/s & s/s] sides of the weld. Keep the locations close to the 'toes' of the weld.

Or get a good welding inspector that is experienced in performing UT on welds. A very good weld is 'clean' and has no signifigcant reflectors in it. A bad weld has significant reflectors. Calibration blocks are to ensure that everybody accepts & rejects the same on poor welds, with some reflectors. You can leave Accept/Reject to somebody with these qualifications. You cannot use 'just another UT technician', though. Techs have to have objective criteria for these decisions.

This 'experience-based' inspection will not meet Code, except on a "Best Effort" basis, so be careful using it.

 

[JR97]

Compit,

In the past we had a similar situation with a stainless steel safe-end to a clad carbon steel nozzle forging. When testing from the carbon steel side the sound just did not want to go into the stainless weld metal. It would deflect down the fusion line and create a false signal at the clad - base metal interface. Unfortunately we repaired a few units chasing these ghosts until we were able to connect the dots. Enough history, what we ended up going to was angle longitudinal waves. The down side is the testing is limited to the first leg. In our case we were able to scan from OD and ID from both sides of the weld to get full coverage. It sounds as though you do not have access to scan from the ID. To get full coverage you will need to grind the weld flush and scan over the weld in addition to from both the carbon and stainless steel base materials.

I agree with Duwe6 that a prototypical calibration / mock-up block should be generated to be sure you are getting the level of inspection you think you are.

Good luck,

JR97

 

[rustbuster]

Yes to needing a calibration block. Consider Phased array and TOFD as well.

 

[Guest102023]

There are advanced ways to do this, including conventional UT. As Duwe6 says, it is possible with an experienced inspector.