Thin wall tube internal or external pressure stress difference on the even thinner end 1

[ThinCap]

We recently had a tube end burst incident. The pressure is from outside of this tube:

The end broke and shot out from the middle of the tube. We usually test it with pressure from inside since when the thickness is uniform, there probably not much difference from either in or outside. And we never had any failure when tested from inside.

But when the end is so thin, quarter thickness of the tube wall, the yield mode/stress will still be the same? I can't find any theoretical solutions for this case. Did any of you had similar situation and analysis before? What is your conclusion? Thanks!

 

[IRstuff]

Well outside my wheelhouse, but one can look at aluminum soda cans and see that both ends, top and bottom, are substantially beefier than the cylinder portion. Moreover, the bottom is concave inward, specifically to resist the internal pressure.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[rb1957]

there are lots of text books on this … Timoshenko "plates and shells".

sorry, but very poorly worded problem statement.

were previous designs with a 0.020" thk end cap ?

"pressure from the outside" would crush this tube, is that the failure that happened ?

previous designs pressurised the inside of the tube, putting the tube in tension ?

external pressure puts the tube in compression, which has much lower allowables than tension.

another day in paradise, or is paradise one day closer ?

 

[ThinCap]

Sorry for the poor...
No, solutions (like the ones in your textbook) are for the same thickness;
No, it only broke the end and pushed it out like a disk;
No, pressure test is usually done from inside (simple), application pressure is from outside;
How is so? Are you talking about buckling.

 

[rb1957]

"pressure test is usually done from inside (simple), application pressure is from outside" … this is very confusing … are you testing the tube with internal pressure (creating tension in the tube) and trying to show the tube good for external pressure (which would create compression in the tube and very different failure modes) ?
generally structures with external pressure are much heavier than for similar internal pressure.
is this the first time you've applied external pressure to this tube ?
Have you assumed that an internal pressure test is showing the tube good for external pressure ??

"Are you talking about buckling" … that is one failure mode for compression. I suspect that with this very thin gauge you'd see crippling and crushing.

"... solutions (like the ones in your textbook) are for the same thickness" … I'd consider this to be a 0.005" thick circular "plate" with a pinned edge.
I'd consider the tube as a separate element.

how on earth did you make the 0.005" thick endcap ? maybe by chem milling ? if so, could that have contributed to the failure ?

another day in paradise, or is paradise one day closer ?

 

[XL83NL]

Wait. You have external pressure, but after the incident, the end cap got pushed out?
Internal and external pressure have different failure modes. It’s time to go back in the basic mechanic text books. You don’t even need ASME to find out what you’re doing wrong.
Why would yield be different when you have a thin end? Is this material completely different from the also ‘thin’) 0.020” wall?
Some very basic fundamental things seem wrong here, or it’s semantics playing a role (are you from a non English speaking country?

 

[ThinCap]

Yes, if you are standing at the open end of this tube, the end cap will be pushed out/shot right at you. Is this good English?

 

[ThinCap]

@rb1957,
There are two separate issues here:
1. This part will be under external pressure in application and it's tested with internal pressure during production only to check leak;
2. The .005" thickness is made by mistake and failed in application.

This is the reason I want to check if there is anything wrong in the whole setup.

 

[hydtools]

Is there internal fluid such that when external pressure is applied the cap failed outward? Like a tube of paste.

Ted

 

[ThinCap]

@hydtools, no, internal is exposed to ambient air.

 

[ThinCap]

If we forget about the tube, only look at the end disk attached someway around the edge: is there any difference stress wise (mode or value) when pressure is external or internal? I can't tell any difference in theory. By depends on how this thing is machined, that edge will make a big difference, right?

 

[rb1957]

it sounds like the base was held, and under external pressure the tube tried to invert itself though the base opening.

I think you can find enough in the texts to analyze the 0.005" base (as a circular plate with pinned edges).

I think the tube sidewall failed as the base tried to invert itself into the tube, putting a lot of bending on the side wall. This will be very hard to analyze (possibly impossible without FE)

another day in paradise, or is paradise one day closer ?

 

[rb1957]

how did you machine 304 SS 0.005" thick ??

another day in paradise, or is paradise one day closer ?

 

[ThinCap]

Yes, around the tube ID edge, the stress should be the highest. But this is true with both external and internal pressure, right?
I think they gun hole drilled the center and lathe the outside, then lost track of the end wall thickness...

 

[rb1957]

true, the end-cap is experiencing a very similar stress state (just opposite faces feel the same stress), but
1) the sides of the tube are very different, and
2) the behaviour of the tube is very different (under internal pressure it wants to expand like a balloon, under external pressure it either crushes itself or tries to invert itself).

I doubt they could machine a 0.005" thick wall if they tried to.

another day in paradise, or is paradise one day closer ?

 

[ThinCap]

I just checked, for 304 SS material, the tensile strength is about double the compression one. So using internal pressure instead of external is not a valid strength test. I already informed test lab about this.
But since I don't have access to a metallurgy lab, I still don't know the cap failure mode: is it by shear, bending, or both?

 

[rb1957]

and the real structure will probably not achieve fcy if that's what you were looking at.

"using internal pressure instead of external is not a valid strength test" … indeed it is not.

another day in paradise, or is paradise one day closer ?

 

[dhengr]

ThinCap:
RE: Internal pressure…, means tension stresses on the interior, at the reentrant corner btwn. the end cap and the cylindrical tube, a very high stress concentration location and likely failure as if an i.d. shaped disc. just unzipped from/around the inner cylinder circumference.
RE: External pressure…, means compression at that internal reentrant corner, not a likely sudden failure condition. There will be very high tension stresses at this same radius, on the outer surface of the end cap.
As mentioned above, the cylindrical tube acts completely differently when under internal or external pressure. In the internal pressure case, the cylinder will yield in tension, a fairly slow failure mode, which can be seen and measured. In the external case, you will almost certainly get a buckling failure of the cylinder, a sudden failure.

You may use internal pressure to check for leaks, but it should be at fairly low pressures; and a timed pressure lose, or bubbles in a water tank type of test. This, so as to limit stress concentrations at the reentrant corner btwn. the end cap and the cylindrical tube. There really is no way that an internal pressure test will represent the real world failure mechanism or ultimate strength of that vessel when it is actually used under external pressure. Whether you can machine a .005” end cap on that tube or not is really questionable. However, it is a certainty that a .0005” or .0001” tool mark in that cap will be a significant stress raiser when related to that thickness, in a high tensile stress region.

 

[ThinCap]

I have the same tool mark theory: end mill cutter tends to leave a mark at the cutter tip, where stress is already the highest. I don't think the end disk has uniform thickness either, usually the middle is thicker.