App. Y Gasket Issues

[david339933]

I have a flange pair designed to Sec.VIII Div. 1 - APP. Y.
Flanges were designed to use an EPDM Gasket classified as self-sealing. Now vessels are fabricated, customer is having issues with degradation of the EPDM in service with vessels previously fabricated (Service: alcohols/trace aldehydes) at design temp. of 356F. I am now in search of a gasket which can sustain the service while maintaining a low load due to seating pressure HG.

I have been in contact with multiple gasket manufacturers who state no elastomer will last under these conditions....and suggested gasket had a y factor of 2500 psi....to great for the flanges.

Any suggestions would be great.....

 

[SnTMan]

david339933,maybe some more info. O-ring style gasket? Failing on bolt area?

Regards,

Mike

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

 

[david339933]

Mike,
Yes, o-ring style groove depth 2mm, width 30mm. When I add HG to Wm1....where Wm1=H + HC + HG, Bolt area Am is almost double Ab....and bolt stress is therefore too great as well.

 

[SnTMan]

Hmmm, not really an O-ring then. Teflon gasket maybe? It would seem higher strength bolts likely not feasible...

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

 

[david339933]

I have the feeling I am missing something....or am so confused I don't know which end is up.

I have used App. 2 for HG where HG = W - H
W = (Am + Ab)Sa/2 (per App. 2 for gasket seating)
I have Am as Am2 per App. 2 where Am2 = Wm2/Sa for gasket seating.

Am2 is small using Am2 = 3.14bGy compared to the actual bolt area Ab. As W is increased to the average of min. required area and actual area, W becomes much greater....resulting in HG being greater....as well as HC as it is HC = (MP + MS)/hC. MP = HDhD + HthT + HGhG

As HG is greater it increases Wm1 (Wm1 = H + HC + HG) in App. Y to the point that Am>Ab....

Not sure you can follow this mess, but if so does it make any sense?

 

[TylerM]

There are elastomers that will operate in an organic solvent environment at the temperature you're using them at, I remember finding one a few years ago but it was very expensive. Just a quick check on McMaster and I found a Viton® Fluoroelastomer that has a hardness of Durometer 55A and can operate up to 395F. (It's a cord stock not a ring, but at least that's a start)

https://www.mcmaster.com/7643k71

You might also want to check into non elastomer rings such as HDPE, LDPE, Teflon or Nylon those materials are chemical resistant and hopefully you can find one that won't stress your bolts too much.

 

[SnTMan]

david339933, a little difficult to follow, but sounds about right. Are you saying that Wm2 is greater than Wm1, and so governs W?

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

 

[david339933]

If I use Wm1 and Wm2 calcs from App. 2, yes Wm2 is over double Wm1.
But using App. Y calc for Wm1 it is much larger as Wm1= H + HC + HG per APP. Y
I am using Wm2 to establish W for gasket seating in App. 2. Then using HG = W(gasket seating) - H and plugging into Wm1 calc.

 

[gasketguru]

I'm not sure I understand the housing here - 2mm deep x 30 wide - are you coming metal-metal on the outside or something, or is this a regular gasket with variable compression in a spigot-recess etc.?(It does not sound like an O-ring for instance....) As you say, if the diameter and pressure increase then operating rather than bolt-up controls the load value, though to be honest the ASME method is flawed anyway as it uses the concept of "effective width" (the sqrt bo/2 part etc.) so you tend to under-estimate the load in many cases anyway.

The duty and diameter are not given - typically rubber gaskets are only low pressure duty (typ. <100 psi) unless you put an O-ring in a groove and come metal-metal. In practical terms a rubber gasket might be loaded to 500-1500 psi on the actual contact area depending on size, thickness hardness etc., and remember that rubber is a fluid and does not compress, it displaces.

In terms of materials your temperature is high for basic EP rubber - I would not offer Viton (Fluorocarbon) as mentioned above for alcohols (presumably methanol, ethanol, iso-propyl etc.) - maybe silicone (though mechanical strength is not great), or perhaps something like "Aflas" (a fluorinated EP) is good at high temperatures. Another option if you need something soft and compressible might be ePTFE (e.g. "Gore-Tex" or similar) as a sheet gasket or even expanded graphite, but again is compression in a fixed depth housing or variable?

 

[david339933]

Yes, Metal to Metal contact. MAWP is 87 psi. Mating flange is FF.

 

[SnTMan]

david339933, given that Apx Y is more or less silent on how Hg is to be calculated I wonder if you couldn't just use Wm2, b*pi*G*y from Apx 2. I don't know if that will get you out of trouble or not. I'd like to know what y value was used for your original EPDM gasket. Was it really that much lower than 2500? Also I am curious of your choice of this modified Apx Y design approach over a standard Apx 2 flange, if you'd care to share.

You and I had discussed the calculation of HG in thread292-435186, it seemed reasonable at the time, and still does I think. However your are in a different situation now. Maybe time to revisit.

I'd second gasketguru's recommendation of the "Gore-Tex" type of product. It seals very easily and solves many field type problems.

Regards,

Mike



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

 

[gasketguru]

The Hg is of course the forces at the gasket reaction, and the App. Y says it is the gasket load due to seating plus the load due to the self sealing of the gasket - thus, whether you are taking the gasket load as being the traditional factor derived value (i.e. the 2b x Pi x G x m x P etc.) or something else, there is also the fact that the rubber is a fluid (Poisson's effectively 0.5), so 87 psi in the vessel means 87 psi in the rubber as an additional pressure load in that location.

I still wonder about the practicality of the rubber gasket here - can you reliably obtain a 3/32" or 2.5mm thick sheet perhaps at this size and have around a 19 or 20mm (?) wide joint in the 30mm recess (even then you have about 90% fill, so a bit close to over-fill if the tolerances are not good). A rough estimate based on loaded vs. force free areas and modulus etc. might give something like 10-15 MPa in practical terms to get 0.4-0.5 mm compression on a gasket of perhaps 732 x 776 dia. - I've not seen that many joints done like this over the years - an O-ring or just a FF gasket tends to be more common.