# High temperature pressure vessel seal

## Main Question or Discussion Point

Hello,
I am looking for a seal design that would seal between the end plate and cylindrical
section of a pressure vessel .
The vessel is made of hastelloy X and has a working pressure of 4000 psi and
a working temperature of 1600 deg F.

-------------

------
!
!

Excuse the old school graphics but the idea is to give you a general idea of what
I am trying to seal.

I have found a graphite sheet material to use in a face to face seal.
http://www.pressureseal.com/graphite_product.pdf [Broken]

Are the record groove cuts in the face necessary to keep the seal from moving.
Is there a formula for calculating how much clamping force would be necessary.

Any help would be appreciated,

Greg

Last edited by a moderator:

Related General Engineering News on Phys.org
I'd be concerned about any gasket type material at 4000psi. You really need a professional pressure vessel designer for this, to avoid killing yourself and others.

Hello, I have talked to several companies that make high pressure vessel and heat exchangers
and have essentially copied there design but i was looking for some more definitive calculations because of the temperatures .
All the crush seal systems I have seen will not survive the temperature.
O rings of the highest temp rating of 750deg wont take the heat either.

Looks like a graphite seal is the only option for the temperature.
With aerospace gradebolts to achieve the required clamping force.

Greg

Don't forget to derate all the yield and ultimate strengths at (and beyond) the planned operating temperature. I'd suggest putting this in a bunker.... I can't advise on seals at this temp - Cu crush gaskets were my first thought, but we're close to the MP. Also make sure any seal is compatible chemically with whatever is inside this container. Find a pro to review the design - even if you think you are just copying someone else.... I'm comfortable doing 1000psi designs at room temp, but certainly not at 1600F+ !

Also, be concerned about the oxidation of any materials at this temperature.

And the large thermal expansion effects between one material and the other. If everything doesn't equalize thermally on startup, you could possibly overstress the bolts. Worry about hot/cold cycling and how that would affect the preload on the bolts.

Hello,
The pressure vessel has a designed working presure of 7000psi and a burst of over 15ksi at
1600deg F and about half that at 1800def F
The only real limitation has been the seal for the endplate because standard crush seals
are near there melting points at this temperature.

Probably going to have to try the ceramic and then subject it to a hydrostatic test.
at the lower temperature the ceramic will fail way before the hastelloy.

Greg

What did the original designer use for the seal? A room temp hydrostatic test would be almost worthless...

Unfortunately your wrong at the fact,
If you know the material yield point at room temperature then you can calculate what it will be at high temp.
That how they test high temp pressure vessels for the chemical industry ,kinda hard to hydrotest above the boiling point of water.
There is an allowance to use gas in the pressure vessel code ,but nobody does it that way.

I have watch them test pressure vessels the size of semitrucks as well as for valves for pipeline.

There is a definite reason why hastelloy was chosen ,is has high temp strength,it is unaffected
by oxidizing or reducing atmospheres.

Greg

Q_Goest
Homework Helper
Gold Member
I'm sure what RocketSci meant was that testing a seal at 4000 psi at ambient temperature won't give you any guarantees that the seal will function properly at 4000 psi and 1600 F. The actual operating conditions will likely be such that some parts of the vessel are much hotter than others. The resulting differential thermal expansions could resulting in higher or lower gasket loading under operating and transient conditions. Also, the gasket material properties such as compressive strength, hardness, etc... will vary with temperature. So testing a gasket at ambient may not give you a good demonstration of gasket worthiness at the higher temperature.

I design a lot of very low temperature, high pressure gaskets so this is a bit of a step out for me as well. If I had to do it, I'd consider using the gasket material you've selected, I'd consider using a crush gasket of some material that won't melt such as a nickel alloy, or use a metal to metal seal. Metal to metal is used for pressures in excess of 100,000 psi. The trick is to have very smooth surfaces and ensure the contact stress under various operating conditions is well above the differential pressure being sealed.

To ensure the contact stress will not be relaxed at operating temperature, you should consider what temperature the bolting and other joint parts are going to be at under various operating conditions. The first operating condition normally considered is the steady state one. In this case, the vessel contents are at 1600 F, but that doesn't mean the bolting or other parts of the joint are going to be at the same temperature. Generally, bolting is thermally isolated to some degree from the vessel and gasket. This is especially true for the second operating condition which is your transient condition. Because the bolts might not be at the same temperature (they're likely going to be cooler in your case) the bolts won't have expanded as much, so they could be under higher stress at the operating condition. If I were designing a seal for the vessel you've described, I'd look to try and insulate the bolts from ambient so that they might warm up to the same degree as the vessel and minimize differential thermal expansion.

Besides using gasket material, a metal to metal (Hastalloy) seal might work. The surfaces should be very smooth, say at least 32 RMS, and contact stress should be close to the yield stress of the material. Also, see the ASME Section VIII, Appendix 2 for ASME code compliance or the applicable pressure code in your country.

Hello,
So your advocating a surface to surface seal vs crushing the graphite seal?

I went with ASTM A193 B5 bolts to give me a good clamping force and plenty of
margin in yield strength of the bolts.
I found a formula for face to face gaskets and it shows that the clamping force
on the gasket must exceed the pressure in the vessel.
This going to require atleast 20ksi of clamping force to have any margin for
thermal expansion.

Are you sure a 32 rms finish is smooth enough for a face to face seal,that is not
smooth enough for test blocks to stick to.
I was thinking maybe a grind finish of a 6rms .

And yes the transient condition does concern me due to the inlets of the pressure
vessel by nature is going to be at a lower temperature that the vessel .

Greg

Q_Goest
Homework Helper
Gold Member
So your advocating a surface to surface seal vs crushing the graphite seal?

Are you sure a 32 rms finish is smooth enough for a face to face seal,that is not
smooth enough for test blocks to stick to.
I was thinking maybe a grind finish of a 6rms .
I wouldn't say I'm advocating the metal to metal surface, but it should work. Look at high pressure, cone and thread tubing for example, which goes to 100,000 psi or even higher. They use stainless steel, work hardened, with a cone face surface finish of around 32. Once it's used, the material on the surface does plastically deform and become even smoother, but I don't think it's important to have a super fine surface finish to start. The surface contact stress is the most important. I'd think a contact stress on the order of 20 ksi is a bit light. It should be close to the yield strength of the materials in contact.

Hello,
The 20 ksi figure i quoted was the bottom number to have any safety factor for thermal
variation/pressure variation/etc the number have would have to be higher.
With 110ksi tensile strength bolts that shouldn't be a problem.

I have worked with tubing in the 50-75ksi range and when everything is perfect it works,
but if there is any misalignment of the connection it leaks.
I have seen guy try to over-tighten the connection to stop the leak and it
just makes it worse.

I see a metal to metal surface working as long as the two surface are parallel and flat.
If the plate try's to potato chip from thermal distortion or the cylinder egg shapes,
there could be a problem.

Greg

My key concern here is that you are proposing a red hot, 4000psi pressure vessel, which I'm not sure has ever been done before. Simply noting that Hastelloy has a workable yield and UTS at this temperature is only step one. Have you considered the creep characteristics as well? What are your safety mitigation plans if this thing actually does fail catastrophically? What pressurized volumes are involved here? Does the entire volume really need to be heated to this temperature, or could you heat only the outflow of this tank? You really need to talk to many ASME pressure vessel designers and hire one who thinks that this can be made safely. The fact you are asking forum questions shows you don't have the required knowledge. (and nor do I for that matter).

Hello,
I find it interesting to see what kind of response I get.
The thing is the pressure vessel has already been built and tested
several months ago.

The things I didn't bother to mention was we used hastelloy bolts.
We used the ceramic gasket with a considerable amount of clamping pressure.
We wrapped the whole assembly with an aerogel blanket to maintain an even
temperature .

For a first test we hydro-tested out to 7000psi at room temp,
Then tested with nitrogen at design pressure at steadily increasing temperature,
with strain gauges and indicators monitoring strain and stretch.

Greg