composite vacuum chamber

[jim lee]

Not a boat, but the same as a submarine. I'd like to build a composite vacuum chamber. For now, two hemispheres radius R with a small return flange to set them together. My question, how does one calculate the lay up necessary to not collapse under one atmosphere of pressure?

Many thanks in advance.

-jim lee

[Eric Sponberg]

You should consult an engineer to apply the appropriate engineering calculations to the task at hand--spherical plate strength and deflection. The thickness-to-diameter will be important, as will the strength of the fibers and resin that you use. The loading is easy--14.7 psi surface pressure. The rest is more difficult, but the appropriate composites and/or mechanical engineer can get you an answer.

Eric

[jim lee]

Well, this is why I posted it. Most likely there will be an engineer around here that understands this kinda' thing.

-jim lee

[waikikin]

Hi Jim, what are you gunna use it for? Regards from Jeff.

[jim lee]

Its just an experiment I've been dying to try for a long time. I wanted to see how light I could make one. But I'd no idea how to calculate the stresses involved or the laminate schedule necessary.

-jim lee

[SamSam]

I have no answers but it would seem the more perfect the sphere, the thinner it could be, as forces would be more evenly distributed. Any flattish areas or lumpy areas would be weak points, and any force from the outside, like being bumped by something such as a corner, would have to be taken into account. Also the size of the 'chamber' would seem to have an effect on the thickness of the walls.

[Rick Willoughby]

Not a boat, but the same as a submarine. I'd like to build a composite vacuum chamber. For now, two hemispheres radius R with a small return flange to set them together. My question, how does one calculate the lay up necessary to not collapse under one atmosphere of pressure?

Many thanks in advance.

-jim lee

Shell stress for thin shell on isotropic material is:
Diameter/ (40 x thickness) in MPa

You will need a triaxial layup of course. It does not need to be very thick because it is not a very high pressure.

Have you ever tried a compression failure on an egg by holding it longways between your palms and trying to break it by squeezing. Will take a huge load if evenly applied. Spheres are very good shapes for handling compression forces.

You will need some margin to allow for lack of ovality and variation in thickness. Any point load for supporting it will also be significant stress raiser.

Rick W

[jim lee]

MPa? Sorry, this isn't my normal stomping ground so I'm not familiar with the units. Guessing 1,000 pascals?

-jim lee

[Rick Willoughby]

MPa? Sorry, this isn't my normal stomping ground so I'm not familiar with the units. Guessing 1,000 pascals?

-jim lee

M means 10e6. So a million pascals. Normal air pressure just happens to be very close to 0.1MPa.

The compressive strength of mild steel is around 400MPa.

Rick W

[jim lee]

I guess I'm still too far out of my pond. I can't do the jump from shell stress to composite layup schedule.

Thanks though..

-jim lee

[Rick Willoughby]

I expect anyone would find a composite layup for a spherical shell to be very complex. It almost needs to be spun as a cloth would inevitably buckle in the layup and this reduce the compressive strength dramatically.

You might get away with narrow tape and wrapping it in such a way to cover the entire surface in a triaxial pattern.

The stress I provided can occur in any direction so you need triaxial layup to get fibres arranged efficiently. Also from my own measurements and test data for composite panels you find that fabric layups fail in compression around half the rated tensile strength. However this is in flat panels or single axis log and the failure is usually caused by buckling. I cannot wrap my brain around how this relates to a sphere because a sphere is an inherently strong shape but I would not exceed half the rated tensile strength before applying any safety factors for stress raisers.

Given the strength of these materials it would not need to be very thick relative to the size of the sphere. As an example take a look at an incandescent light globe. These are under vacuum.

Rick W

[Rick Willoughby]

I thought of another common example of a sphere under pressure. Or more accurately part of a sphere. Have a look at the bottom of a common pressure pack can. You will see that it is hemispherical. If it was a flat bottom of the same thickness it would belly out and most likely tear at the seam. Being hemispherical it has much low stress than if it was flat.

Rick W

[jim lee]

I think I'll try it from the other end. I can calculate the lbs/sq ft I can allow for the weight of the skin. I'll come up with a composite that's a tad less and just see if it works.

-jim lee

[Tarek]

Just out of curiosity, Jim. What radius are you considering?

[rxcomposite]

Jim,

Rick is correct and is leading you the right way. But if i were to do it, I would filament wind it and go more than triaxial. I assume you are building a spherical shape so all stresses are in multiple direction. Using filament winding will eliminate fabric overlap.

I would say all stresses on the "skin" or plate is compressive in all direction with just a little bit of shear in the neutral axis. The inner part skin, with a lessr radius than the outside will compress more. The difference of contraction with the outside layer and the inner layer will cause it to slide past each other and cause horizontal shear.

I guess you will have to learn composite to learn the lay up schedule or you can experiment by progressively building up the layer while continously increasing the vacuum at as you build up (and cure) after each layer.

Rx

[jim lee]

I'm thinking 3' radius to start. If I can get the composite to weigh in less than 1 lb/sq foot, my calculations show, it'll float.

In air.

-jim lee

[Rick Willoughby]

I'm thinking 3' radius to start. If I can get the composite to weigh in less than 1 lb/sq foot, my calculations show, it'll float.

In air.

-jim lee

The 1lb/sq.ft is OK for a layup having sufficient robustness to take the local loading and allow a triaxial arrangement of weave.

My calculations show you need to go much bigger with a skin that heavy. Post your calcs.

Rick W

[jim lee]

Weight of skin < weight of air

4 pi r^2 x (skin lbs/sq ft) < 4/3 pi r^3 x (air lbs/cbe ft)

I don't remember weight of air off the top of me head. Give me a few.. To look it up again..

-jim lee

[Rick Willoughby]

If you google vacuum balloons you will find some interesting stuff.

From a practical point of view I think you would be much better off using a cylinder. You will find forming a sphere to be very difficult.

You should be able to do a single layup of glass cloth into a cylinder that is not too delicate to handle that has a skin weight around 300g/sq.m. It only has to handle the hoop stress from 100kPa. The ends could be domed, strutted or thick panels with light core.

The attached pictures shows a light structure that almost floated before it collapsed. It is not the best shape from a stress point of view but shows an alternate of using struts rather than relying on the skin stress.

In any case it is possible to make a vacuum balloon that has some lifting power.

Rick W

[jim lee]

Well, here I am surrounded by fiberglass infusion equipment. So I figured all I need to do is make a big hemisphere plug -> Mold. Pop out some big 'ol hemispheres, glue 'em together and see if it worked.

So, I looked at a bunch of the "will it work" stuff and pretty much everyone says it won't. Odd, I have to go back over my math, but I think it would.

I'm surprised that cube thing worked at all.

-jim lee

[Rick Willoughby]

How will you get a spherical layup that is not creased? Any creasing will create a local buckling failure.

Rick W

[jim lee]

Creases? I don't see that as a problem. I figure it'll be just a big 'ol infused hull shape that happens to be real close to a sphere. I plan on doing it as a cored composite. That should give some poking resistance.

From what I've been reading, I'm wondering if I have a decimal point off somewhere. Seems no one has been able to do this. I'm going to go through the numbers again before trying it.

-jim lee

[Rick Willoughby]

Typically hulls do not have continuous compound curves like a sphere. That is why I suggested a cylinder. I imagine you could layup about a dozed pieces with overlap.

For a 3ft radius you will need a layup less than 1.3oz per sq.ft for it to be buoyant at room temperature. Maybe possible.

Rick W

[jim lee]

Really? How big does the radius get when the layup is 1 lb / sq ft?

[Rick Willoughby]

13m radius or in your British units say 80ft in diameter. It will weigh 11 tonne before the vacuum is pulled.

It is not a trivial task. If it was you would expect to see a lot of vacuum lifters around rather than cranes.

Rick W

[jim lee]

Yikes! looks like I did move a decimal place over one somewhere.

Rats!

-jim lee

[Rick Willoughby]

You should check my calculation.

I also think a much lighter layup is possible so the 6' diameter might be practical but I do not think it will be done with infusion. It will require something like 100gsm cloth (more like film than cloth) either side of a very low density core.

A 6' diameter ball would be around 10lb to get buoyant.

Rick W

[rxcomposite]

Jim,

Actually, it is a little bit of "rocket science". I programmed the calculations for Rocket Motors and Pressure Vessels for you.

I reversed the pressure and used compression instead of tension for the stress.

I am attaching the excell program so you can play around with it. Seems the 14.7 psi vacuum will require only 0.001" of skin so the problem would be how to support the skin since it will collapse unto itself. Finding a fiber that is less than 0.001" is also impossible.

The netting analysis is an entry level method for predicting the sizing of materials. No interlaminar shear, no stiffness analysis.

You gave me an idea. I can use this to design submarines.

Rx

[jim lee]

Well, to solve the collapsing problem I was going to core the thing. This would also go a long way for solving the "It implodes if it touches anything" problem as well. I have a bunch of 1/4" core left over from the Dart Project.

From what you are saying, it seems one would need very little glass to hold the pressure as a sphere. So mostly the deal is the weight of the core.

-jim lee

[rxcomposite]

yes Jim,

Stiffness would be the problem and cored composites is the best approach.

In cored laminates, the skin receives the highest stress, the core receives compression and shear.

Try also increasing the safety factor. Aerospace generally use 2, Marine uses 3.

Rx