Vacuum chamber flat end-caps

[drodrig]

Hi there,

I have to design a vacuum chamber, big in diameter (~2.5m in diameter). On one end I have to use a flat cap, as thin as possible.

My first challenge is how to compute the stresses in this end-cap. I learnt how to calculate deposits with round endcaps (hoop stresses...). But how does one check this one it is flat?

thanks
regards,

 

[TGS4]

For vacuum (external pressure), the stresses are secondary to knowing that the failure mode is buckling. The appropriate question is how do you calculate buckling of a large flat plate under external pressure.

These types of chambers are generally designed to a know world-wide Code.

 

[LittleInch]

What shape is your chamber? If it's circular you will find some data, but thickness usually ends up far thicker and heavier than a hemispherical or toroidal head.

At 2.5m you'll probably need some radial reinforcement. It's a complex model?, especially at the connection to the chamber. You might need FEA.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[rb1957]

thin flat plates are lousy for carrying compression. if you're using flat plate, use many stiffeners ! ... 8ft diameter, 15psi ... lots of bending !

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

 

[EdStainless]

If you really want it thin pay attention to what they said above.
It will either need internal bracing or it needs to be domed.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[drawoh]

drodrig,

Why does it need to be thin? Will there be room for gussets?

A thin wall requires accurate design analysis, high quality fabrication, and all sorts of inspection during fabrication, and after installation. You are adding cost to your design and fabrication. I hope you are adding value somehow.

--
JHG

 

[racookpe1978]

Fatigue and vibration going to be a problem (like in a space launch or while the internals are loaded/unloaded) or will it "safely" sit in a lab environment with a static load?

What's the penalty (safety and cost) if it (well, when) does fail catastrophically if your stress model is wrong?

 

[drodrig]

Hi there,

For now I am interested about the end-cap, the cylinder will come later. What code?

Bucking is important, but on the other hand we have "only" 1 bar of pressure difference. In principle we have just a cylinder with flanges and the endcap. I tried some FEA, but since the cap is so thin (I used 1mm) wrt the diameter it is not easy.

In principle bending is not a big problem, the problem is it should not break

It needs to be thin because particles go though it and we want to "see" them, if the material is thicker particles get "broken" in the material. ribs/bracing/stiffeners are also bad (more material)

No problems with vibrations or fatigue, it will be placed on a hall.

Here is a picture of something similar, but in this case they used the domed geometry.

http://cds.cern.ch/record/40739?ln=en

What I would like to know now is how to calculate a circle; clamped on the edge with 1 bar on one side. I can calculate beams (one dimension way bigger than the other two) but a flat surface...

 

[EdStainless]

The only way to keep it flat is to hold it under very high tension.
Link they do blades for ID saws (look at page 10

http://www.winter-superabrasives.co...nd-ID-saw-blades-semiconductor-industry_e.pdf

) I have also seen systems where they clamp them flat then use hydraulic pressure is a groove to tension the blade.
If you are using a plastic film then perhaps stretching the outer edge over an o-ring would allow you to tension and seal it.
It is going to take some very sturdy flanges to carry the load without distortion.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[LittleInch]

It's called a blind flange.

Work out the load. It might only be 1 bar, but it's a big area and works out to be 50 tonnes force.

That's a lot for 1mm to try and hold back. ...

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[LittleInch]

The code is probably ASME VIII.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[drawoh]

drodrig,

Is the transparency of your end a function of thickness, or is mass or some other material characteristic important too?

--
JHG

 

[rb1957]

"In principle bending is not a big problem, the problem is it should not break" ... not in my world ! btw, what's "thin" to you 1mm thick ? 1cm thick ??

The main problem you have is that the vacuum tank puts the shell in compression, unlike a balloon skin.

"It needs to be thin because particles go though it and we want to "see" them," ... particles, what, like sub-atomic particles ? we are talking steel for the endcap material, right ?

I think your best approach is to dome the endcap somewhat. If you dome it outward, you may get a somewhat satisfying (if really disconcerting) span-through ! It should minimise the amount of stiffening.

It may be "only" 1 bar pressure to you, but to me it's an enormous load ... 2.5m dia = 98in (call it 100in), near enough 8000 sq.in, over 115,000 lbs, over 50 tons (or tonnes)
from Roark, max. moment = 3.3/16*14.7*50^2 = 7600in.lbs; 1in thick, stress = 46ksi

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

 

[Compositepro]

You will never have a thin cap that stays flat under such high load. The best you can do to keep the thickness low is similar to the CERN design which you linked. Kevlar is an organic material with very high strength, which only contains carbon, hydrogen and nitrogen. Carbon fiber is stiffer and stronger but more fragile. Most people would call this a pressure bulkhead. It will need a thick, strong flange to support the tension in the pressure membrane. Carbon fiber is used in x-ray tables and in nuclear reactors and accelerators because it will not become radioactive when irradiated by neutrons.

 

[racookpe1978]

OK, so design your pressure vessel (a true, actual pressure vessel big enough to hold the sensor and its support AND the thin low-pressure membrane. Mount the pressure vessel around the end of your accelerator/particle target with the large stiff heavy flanges needed to keep everything rigid.
Then your 1 bar vaccum pressure - that "minor" 50 ton load is on the outside of the PV, the target and sensor are on the inside of the PV, and the thin membrane needs to maintain the pressure diffierence between the accelerator interior (a "pure" vacuum) and the sensor pressure vessel, a "dirty vacuum with an access port.)

 

[IFRs]

Can you forget about it being flat. Does that matter? Can you tolerate some deformation? If you can allow it to dish inwards a bit, it will be in tension and you will need to design the cylinder to take that tension.

 

[SnTMan]

drodrig, recommend a Roarkes for some scoping calculations.

Regards,

Mike


The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[drodrig]

Hi there,

I talked to a colleague today, there will be an inner hole with a flange in the middle, with a diameter of 1000mm (~3ft) and the outer would be 2400mm (~8ft).

3 materials will be considered: aluminium foil, kevlar and kapton foil.

Yes, subatomic particles will go through. The "transparency" depends on the radiation length (let's say not dense materials and with low atomic number)

I was thinking about 1mm thick. The picture of the kevlar dome is .85mm

rb1957, could you point out to the formula in Roark where you get those numbers?

I got the book today and found very long formulas for calculations, I'll put them in excel tomorrow

We need to make it flat so the particles trajectory through this cap is as short as possible. With the dome idea they won't go perpendicular to the surface (not shortest distance)

racookpe1978, I can't put the detector inside the vacuum chamber. There is a big solenoid around and other detectors (which should be in a helium atmosphere). Also maintenance would be harder in vacuum.

Thank you all!

 

[prex]

You must forget to have a 1 mm wall fully flat with 1 bar pressure delta (as pointed out by others above). Your only option is an inwardly domed end.

prex
[URL unfurl="true"]http://www.xcalcs.com[/url] : Online engineering calculations
[URL unfurl="true"]http://www.megamag.it[/url] : Magnetic brakes and launchers for fun rides
[URL unfurl="true"]http://www.levitans.com[/url] : Air bearing pads

 

[LittleInch]

drodrig,

would you not think that with all the resources available to CERN that they would have used a very thin flat plate if they could??

The dome shape is very strong - think of an egg shell - about your 1mm thick - It will withstand large forces when pressed from the pointy end, but direct that force at 90 degrees to the shell and lo and behold it cracks!

What you're trying to do is physically impossible. You will need to change something, be it the pressure, the thickness or the shape of your end connection.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.