Calculating Safe Wall Thickness of a Vacuum Tube

4mm should be plenty. 14.7 PSI is just atmospheric pressure at sea level.

Check some data sheets.

This tank failed at 10.5PSI.

Railroad Tanker Car Crushed! - The Science Classroom

10.5 psi negative pressure is about 21.4 " Hg vacuum. Our condensers withstand 25-26" Hg on a routine basis. The problem may be this rail car had a fatal flaw somewhere.

Mythbusters tried this same trick years ago, got no collapse at all, until they dropped a 4000 Lb weight (concrete) on the tanker car at a slightly oblique angle, producing a diagonal dent in the upper surface. Then it was no problem to collapse.

OR, this could have been someone's newly designed tanker car, a bit short on material as a weight saver, and voila' it becomes a "flat" car.

You are correct. Thanks.

Yes, and that is not safe from implosion. Just look at it wrong, and it collapses, no harm to anything but that part, though.

Seems like the NASA stuff is about buckling under axial loading, not due to hoop compression. Are we not talking about hoop compression, and some change in the material from isotropic to non-isotropic, or even no change in the material, except that some inelastic perturbation could result in a complete collapse of the cylinder wall?

I was thinking section 4.2.3 might address hydrostatic compression.

Here is a better paper on it...

"Prediction of vacuum-induced buckling pressures of thin-walled cylinders"

http://rennes.ucc.ie/~bill/repository/2012-TWS-55_1-10.pdf

If I were the OP, I would get a few cylinders built and tested. There is a complicated formula for the critical pressure for a perfect cylinder, but any imperfections might severely weaken it. I would err on the side of safety and make the walls a little too thick.

After reading (finally) the section you were referring to, I raise my white flag of surrender, and defer to your obviously superior knowledge of the subject, with minor protest about this mysterious γ term (fudge factor) between "theory" and experiment.

I suspect this to be another one of those nonlinear dynamic situations in real life where nothing is actually a "perfectly" isotropic cylinder being buckled to an oval shape. From what I can see, buckling would almost never produce a regular shape of an isotropic cylinder, or a non-isotropic cylinder either. Also a lot depends on aspect ratio of the cylinder, as the NASA paper mentions. Good stuff, I don't think they theory could be improved much, other than to start out with some object with its imperfections and pimples in the model in a super computer, then do finite element analysis?

I put the figures in Eq. (1) of the link, and it comes to buckling pressure about 30bar for worst case number of nodes n = 5, in reasonable agreement with my figure from Roark. Roark does not mention number of nodes explicitly.

As a check, for dia and length both 750mm (same 2mm thickness) the link gives minimum 1.9bar (7 nodes), Roark ~ 1.7bar, so there seems to be fair consistency.

I think proposed design is OK, but OP seems to have disappeared. He didn't say how the vessel ends are arranged - flat plates, dished ends etc, and needs to ensure the shell is held circular.

OP done gone and left.

"I am designing a component to be used under vacuum condition"

As a designer you're not doing very well are you?

With a lot more information and a sizable fee I'm sure one of our members will design it for you.

Thanks for the very positive comment, I was under the impression this is what this community is for?

I am designing a large system not a simple tube, this part is a bit tricky hence the need for some guidance.

Anyway, if this is a typical reply from this site I will now delete my account.

Thanks for being an upstanding member of this community and goodbye.

If you're that thin skinned, you'll never make it....

4mm oughta be enough....

Not as a tube either. He'll implode.

Take a chill pill, tim darby. You don't want your tube to go imploding now do you?

Even a single dent in that cylinder, and my money say she goes flat as a pancake in Belgium.

You need more metal, probably more like 6.5-7mm. On the other hand, if you can assure yourself that there is no way this cylinder ever encounters a sharp blow from some heavy object, sufficient to produce a visible dent, then, the chances of implosion are somewhat (maybe 90%) less.

Have you considered experimenting with this to see what works and what does not work? Or ask OEM suppliers if their vacuum chambers will withstand that pressure differential.

Gone after only two posts. He lasted twice as long as most of the people that ask us to do their design work for them.

That is the problem with trying to make hope float on nothing. Hope floats on helium.

Yes, it can, but not always. If dented, it will fail.

You could probably go down to 2.1 mm and still be ok...

http://www.pchemlabs.com/product.asp?pid=6140

...most of the time, until you dent the tube, then it's game over.

The equivalent at 10" (250 mm) would be 0.166" wall thickness ~~4.1 mm.

That's .083" (2 mm) for 5 inch OD. For 10 inch OD, I'm pretty sure it needs to be scaled up to 4 mm.

But OP needs to not worry about calculating the thickness and just order a 10" OD from those guys!

(http://www.pchemlabs.com/product.asp?pid=6140)

I just checked one of our rigs, 10" diameter (254mm) x 180mm long, 2mm wall.

OP has left the building.

Probably went to see his shrink.

I was wondering what the OP meant by the term "safe". Safe from implosion when not disturbed? Safe from implosion or leaking when a small coin is dropped a meter onto the cylinder? Safe when dropped a meter onto a hard surface? Safe from an implosion for any reason causing dangerous shrapnel or jagged tares?

Safe, as in, I am in my padded cell now, with a foam rubber covering over my entire body safe, most likely.

The technique is to take a theoretical slice along a diameter of the cylinder and do a force balance along that plane, given the strength of the material of construction; the minimum thickness drops out of that calculation, and the safety factor is then the multiplier of the theoretical minimum wall thickness. This method is taught at schools in the UK in Physics classes.

An alternative is to take a piece of tube that is already suitable for full vacuum use, and use that without calculation.

If you're still about, according to my old version of Roark, if yours were a long cylinder (which it isn't) buckling pressure (out-in differential) is just over 2bar.

For a cylinder 250mm long it's 26.9bar.

Normal to use a safety factor of something like 3 on Roark calculated figures, but above suggests you'll be OK.

If you make it 3mm thick, corresponding figures are 7 and 74bar.

What if it becomes dented? Pray, tell us what becomes of the small gecko in the vacuum chamber?

Why would it become dented? I think it's still OK with a moderate dent, but unless we're told it's somewhere vulnerable I don't see why we should consider that. An old-fashioned vacuum tube TV is dangerous if broken, but nobody worried about sitting in front of it with their kids!

Not sure where the gecko came from, but he'll be dead in full vacuum

OP did not mention if this item is exposed to potential external damage (if something could drop on it, etc. He seems worried about space limitations, so it might be that whoever puts the thing together could damage the tube by scraping or banging into it with a wrench. At this point, who knows or cares?

One thing that has not been mentioned is that these tubes, whatever size, should be seamless, no welded tubes.

I agree, since a weld represents a whole line of non-isotropic structure to the vessel as such, could be stronger, most likely will be the starting point for buckling to lateral stresses.

Another interesting side note: Although I had calculated that Lexan meter covers should be able to take air pressure at room temperature stresses to 100 psig (and others thought so just be W.A.G. method and beverage container videos), impingement of any solid object against a flange part that could concentrate stress should be avoided.

I had one of these Lexan meter covers blow up totally to numerous pieces (I was a mere two feet away with back turned, fortunately no shrapnel injury, but a very loud pop preceded about 2 seconds by a sharp ping (the first crack at the impingement is my guess)) at only 30 psig, and yes I know how much force that is on a 6.5" piece of Lexan about 1/16" thick.

Needless to say, I took steps after that to eliminate points of impingement, and spread out loading of any sleeves, seals, clamps, etc. All pressure tests are done now, but de-rated (for my own "safety") to 20 psig. Good to go.

Next trick is aluminum thin-walled bottle formerly containing Automotive style refrigerant refill, drilled and fitted with 1/4" NPT brass fittings that are epoxied in place. I will not subject this to more than the same 20 psig. I am concerned, though, that ripping of this cylinder bottle will take place at the points where drilled for the pipe fittings, even with the epoxy. I think I will tape this bottle with fiber tape before any further testing.

This thing sounds more like a tub than a tube.

which thing is that? Are you (anonymously) so spineless you need a tub to be contained in?

The OP is dead and the carcass has been picked clean to the bone.