Hello Stressers!!!

I have read this article regarding a Large Diameter Steel Pipe Reinforced by Stiffening Rings due to some vacuum conditions. Actually from the article what happen is some cracking occur in welds of Stiffening rings welded directly to pipe due to thermal lag between pipe and stiffening ring (low cycle thermal fatigue phenomena). Line is subjected to a very high temperature above creep range.

Anyway I'm curious regarding the recommendation to fix the cracking of welds due to low cycle thermal fatigue (thermal gradient phenomena). To make it short, the recommendation is to put a gap between pipe and stiffening ring and do not directly weld the stiffening ring to the pipe. Yes, this will solve the thermal gradient problem. My question is from the vacuum standpoint, on how it will restraint the pipe to avoid collapse due to external pressure if the stiffening ring is not welded to the pipe. Maybe I am missing something from the detail of how the stiffening ring is being designed to cater both the thermal and vacuum problem at the same time.

Anybody has encountered this type of detail. Kindly share your ideas. Maybe my imagination is not so deep enough to come up with that detail. The article doesn't show the final detail only some description which clearly mention not to weld the stiffening ring to the pipe.

The reason why the proposal will work is that although the pipe will "shrink" as it attempts to collapse in on itself, it won't collapse uniformly - it'll ovalize as it attempts to collapse, and the stiffener ring will prevent that ovalization.

However, let's go down your suggested route.

First, I would attempt to engineer around the problem. Can we eliminate the vacuum by controlling the process? If not, can we add a vacuum relief valve, with a source of suitable fluid to replace that which is lost?

Second, let's assume we can't, then what you need to do is engineer a better stiffener ring.

Such a ring needs to be thicker in wall, but must have a groove cut into the inside. If you take a cross sectional view of it, it'll look like a U. (Such a ring will also require a weep hole, which should be oriented straight down so as to not hold water.)

This will add flexibility in the axial direction, so the pipe is allowed to grow axially faster than the ring itself. The ring will however continue to have a stiffening effect if the pipe attempts to collapse in on itself.

How loose are these rings? Perhaps the rings have lesser thermal strain so as the pipe heats up and grows radially, the contact is achieved to provide the additional strength. But I wonder if you can depend on it.

Deleted - as evidently my first response made it in.

Thanks Michael & Dave for that info.

I was stuck in thinking that if the stiffening ring is not welded to the pipe the pipe will collapse. But Michael's idea of the pipe not shrinking uniformly but ovalizing I think make sense since some parts will expand which will compensate to the shrinking part therefore it will try to maintain it's shape (this is shown in the buckling mode shapes when running buckling analysis). I think the gap being specified is just enough so that when the pipe expand radially it will touch the stiffening ring.

Another idea I've come up with is to insulate the pipe together with stiffener ring so that the thermal distribution is uniform and local thermal stresses is avoided.
Based from the description in the article, the large diameter pipe is not insulated since thermal bowing was also experienced, as stated in the article some dust were collected at the top of pipe which acts as insulator while the bottom of pipe has no insulation, so some thermal gradient at the bottom & top pipe was developed.

Michael's idea of putting a groove is also viable, since as stated in the article large axial thermal stresses was developed in the weld due to atmospheric cooling at the stiffening ring while the pipe is very hot. The stiffener ring is indeed very wide with 260mm width as described in the article and welded on both sides. So putting a groove will solve this problem since it will not be so stiff in the axial direction. FEA analysis can be performed considering some heat transfer analysis to obtain the temperature distribution between pipe and ring and stress calculated using ASME VIII Div. 2 Stress Categorization approach.

Cheers!!