Expansion Versus Weld Repair

[yamoffathoo]

1978 TEMA C recommended that for 300 psig+ and 350 degF + heat exchangers with 5/8" dia and larger tubes you should either expand the tubes over 1-1/4" min. length in a two-grooved tubesheet or use full strength seal welds.

If the 5/8" dia tubes are presently rolled 3" long in a 6" thick tubesheet without grooves and SSC has cracked through the tubes next to the seal welds, is there a case for expanding the 3" un-rolled section rather than re-welding the seal weld?

 

[metengr]

If the 5/8" dia tubes are presently rolled 3" long in a 6" thick tubesheet without grooves and SSC has cracked through the tubes next to the seal welds, is there a case for expanding the 3" un-rolled section rather than re-welding the seal weld?

No. Roll expansion mechanically locks the tube into the tubesheet. Seal welding provides the leak tight boundary. If you attempt to roll expand the remaining three inches with no seal weld repair, you may have roll joint leaks over time.

During thermal cycling in service, the roll joint can relax resulting in tube roll joint leaks. The seal weld is necessary at high pressures and temperatures to ensure a leak tight joint. I am curious as to how stress corrosion cracking was determined. Are you sure it was not fatigue crack propagation? What is the tube material and service conditions?

 

[yamoffathoo]

Thermal fatigue is exactly what I think is cracking the tubes - I have tried to explain this in my "Heat Exchanger Tube Failure Mechanism" thread.

I was trying to "re-package" the problem statement to draw out more opinions, like yours.

If you could take a moment and read my problem description in the other thread, it should answer all of your questions.

I am convinced that re-rolling should eliminate the tubesheet/tube thermal gradient problem, making seal welding redundant.

 

[vpl]

For interested readers, the post to which Yamoffathoo refers is thread794-278257. You might also want to look at thread378-278970.

Patricia Lougheed

******

Please see FAQ731-376: Eng-Tips.com Forum Policies for tips on how to make the best use of the Eng-Tips Forums.

 

[yamoffathoo]

Patricia, I am flattered but confused, what do cryogenic flow assumptions have in common with heat exchanger tube bundle thermal fatigue?

 

[MikeMet]

I think you may have a case for rolling the remainder of the tube sheet thickness. Years ago Nooter Corp did a study on expanded tubes into a thick tube sheet with no rolling grooves and found they acheived leak tightness and strength equal to welded tube sheet. Of course tube hole and tube OD prep are were critical. If the tube holes was properly prepared and the tube OD in the roll area was in good condition, you can get a leak tight seal that will resist loosening during thermal cycling. I also think you would be better served by hydraulic expansion instead of roller expansion.

Another option if you truly have SCC is to stress relieve the tube to tube sheet joint. I have done this on several occasions, including to prevent SCC of zirconium tubes.

 

[vpl]

yamoffathoo

You posted both around the same time so I made the apparently erroneous assumption that they were related -- part of a bigger problem, if you will.

Patricia Lougheed

******

Please see FAQ731-376: Eng-Tips.com Forum Policies for tips on how to make the best use of the Eng-Tips Forums.

 

[yamoffathoo]

MikeMet,

Sorry for the late reply, I didn't see your reference to the Nooter Corp study - I will look into that, thanks.

The annular gap may be filled with water, which would defeat uniform hydraulic expansion and cause ripples, unless it was incrementally applied in order to squeeze the water past the upper or lower rolled joints.

 

[Duwe6]

Try rolling. It is cheap [relatively] and will not cause any more problems. As you said, the rolling needs to be done in several stages, with the first being 'gentle' to squeeze out any trapped water. Heating the tubesheet to 250-300 deg-F would also dry out the crevice.

 

[yamoffathoo]

Duwe6, Tubes with through-wall cracks adjacent to the fillet welds could be rolled up towards the primary (high pressure) 1-1/2" long rolled joint and a visual check for water made. Tubes without cracks would necessarily need to be rolled 'down' towards the secondary (low pressure) 1-1/2" long rolled joint. I suspect a consistent direction is warranted so it will likely be down since all tubes are not cracked. Down would also reduce the possibility of tube buckling if the cracks close while rolling.

I suspect a final, incremental roll from top to bottom or vise-versa would be required to smooth out the 'squeeze path' or line of partially buckled tube where the water made it's escape. It has been suggested that the 5/8" OD x .0445" thick Inconel 800 tubes shouldn't be expanded more than 7% total (wall thickness reduction).