I'm having trouble understanding hydrostatic pressure - please help? 1

[nanobot29]

Good afternoon,

I'm having some trouble understanding how parts will behave when submerged under 3,000 psi of water pressure. One case in particular - if there's a annular disk and hub assembly (both steel) with an interference fit of 2,000 psi. so my question is: even though the interference fit pressure is less than the water pressure, will water seep through and into the interference surface? my first thought was this this is hydrostatic condition therefore there wouldn't be any deformation that would allow water to seep through. am i thinking about this the wrong way? is there any scenario that would cause water to seep through?

thanks and regards,

 

[desertfox]

Hi
Can you provide more information for example orientation of joint under water,is the joint held at this depth in water by an external force or is it in balance at this depth of water with the mass of the joint?
I think some water could get in the joint,imagine a cylindrical perfect interference fit equal distortion all the way round, now machine a small flat on one of the surfaces, you would still have an interference fit however the flat would allow fluid into the joint.

 

[LittleInch]

It would be good to have a sketch of your system, but from what you've described, I don't think you have an issue so long as three is no pressure differential between one side of the hub and the other. As desert fox says there is always the risk of water in the fit including any voids in the hub assembly which would cause this pressure difference.

This is pretty deep water you're talking about so would seem to be important to get it right, but hydrostatic pressure becomes important when there is a pressure difference.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[rb1957]

i don't think we can offer usefull help without a sketch.

without a sketch the best you can hope for is motherhood, the worse sarcasm.

Quando Omni Flunkus Moritati

 

[Compositepro]

You have to be careful about equating hydrostatic pressure with stress in a material. Although the units are the same they are not identical concepts, and different rules apply. Stress at a point can vary by direction.

 

[Cockroach]

I think the answer is yes and no.

If the interference fit is 2000 psi with a 3000 psi stagnation pressure to water head, then the lid will pop off if the line of action forces the lid off it's bearing face. Otherwise it won't if it acts in a manner to force the lid into the bearing face, thus making a tighter seal.

Sounds like a homework assignment.

Regards,
Cockroach

 

[nanobot29]

Thank you for the feedback. its not a homework assignment. I work for a company that designs large electric motors for surface and subsea applications. for the last 2 years I've worked on design and analysis of the rotating components in the surface motors (no water pressure). I'm being moved into the sub sea applications and I'm trying to figure out how i would approach some of the stress analysis problems with the presence of water pressure. Typically these motors are submerged in water where pressure is 3000 psi. The motors are also flooded internally with water/glycol with a pressure of 3000 psi as well.

The first example i thought of was a press-fit of a disk onto a spinning rotor. This is a calculation that I'm comfortable doing.. i know how to calculate the radial and hoop stresses as well as the amount of pre-load loss due to the centrifugal forces. however, im trying to figure out how the presence of water pressure (on all sides) would or should change my approach.

In general, i'm trying to learn how to approach these problem. i havent started doing any work on the subsea applications but i wanted to get a head start and really understand the physics behind it. if you guys know of any good books out there or articles i can purchase, that info would be a big help as well.

the initial question i posted was hypothetical... if water pressure is higher than the interface pressure of the rotor and disk interface, will water contact the interface? I'll attach sketch shortly

 

[nanobot29]

sketch attached

 

[LittleInch]

If you read the posts above I think were all in alignment that its when differential pressure appears that you have an issue. If its all at the same pressure then its like it is at the surface, only with lots of water around you.

In your example I can't see why water depth would be an issue.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[Cockroach]

You treat your containment as a pressure vessel, most likely the thick wall model. The criteria here is that the ratio of inside diameter to wall thickness is greater than twenty (20). In that case, Thin Wall Pressure Vessel theory will give you less than 5% error; otherwise you must use Thick Wall theory for the case of BOTH internal and external pressure loads.

Your case is particularly interesting to me. I have just completed analysis on such a case resulting from collapse of a threaded joint in a downhole oilfield application regarding a frac application. The reservoir pressure was 4100 psi (28.3 MPa) with tubing string pressure roughly that of stagnation pressure of a brine solution with head of 1800 m. The joint collapsed from buckling of an external load, very similar to your case, except you have a flange.

Not the first time I've seen this or had projects relating to the phenonema. But over the industry, the usage of incorrect models to predict pressure vessel stress is particularly poor. So I decided to get some testing completed on a pressure vessel stringed with strain guages and complete a study for a white paper on the subject.

So I recommend Timoshenko - McCullough as classical literature on the subject. There are some excellent resource material on the subject regarding Von Mises with application to the Hencky model. I would start there and be careful to derive your model from first principles. It's very easy to misrepresent the triaxial state of stress subjected to external loads other than internal/external pressure. Lubricators for example are subject to axial compression from guide wires stabilizing what can be regarded as a Euler Column. Without FEA or similar type of analysis BACKED UP by an independent model for validation of output....well....God help us all.....

Good luck with it. Sounds like an exciting challenge!

Regards,
Cockroach

 

[racookpe1978]

Well, your interference fit will be a "pressure" of 2000 psi ONLY over the parts of the joint that interfere with each other. if you have (for example) a tiny scratch 10/1000 x 10/1000 through the interference fit zone of a 12 inch dia vessel dia x 1/2 wide zone, then your 3000 psi water WILL find its way through that scratch into the interior, despite the 2000 psi x 3.14 x 12 x 1/2 force holding the lid on. (Above assumes your 2000 psi interference fit is distributed over the area, somewhat simplistic, but only valid as an example. An actual interference fit will tend to "close up" small scratches since it is, after all, a overlap of the two surfaces)

 

[LittleInch]

nanobot29, You have some good advice here, but looking at your sketch - which appears to be a flooded motor - it would appear to me that you don't have a pressure difference accross your interference fit and therefore, IMO, the water depth is irrelevant for this particular scenario. If you have a situation where you have atmosphric pressure one side and hydrostatic pressure the other, then yes it becomes a significant issue and all sorts of things need to be considered.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way