Spiral Wound Gasket with External Pressure 1

[JJPellin]

I am investigating a problem with a possible internal leak across a spiral wound gasket. After digging deeper, I found that the spiral wound gasket is oriented such that the pressure in on the outside. I realized that I could not recall ever seeing a spiral wound gasket with OD pressure. I attempted to research this on-line, but did not find any relevant information. I am hoping someone here has some experience in this area.

Just to clarify, I am referring to a standard spiral would gasket of the same configuration as a typical flange gasket for a raised face pipe flange or a head gasket in an API pump. In cross-section, the gasket is composed of wraps of 316 SS with a chevron cross-section. The filler between the wraps is flexible graphite composite. Normally, the pressure would be on the inside of this type of gasket such that the pressure would tend to spread out the chevron, forcing the edges harder against the sealing surfaces on either side. It seems to me that applying external pressure would tend to collapse the chevron shape inward and cause the sides to pull away from the sealing surfaces. But, this is just speculation on my part.

In my particular application, the gasket is subjected to approximately 2000 psid of 350 F gasoil.

Johnny Pellin

 

[racookpe1978]

How much differential pressure? 2000 psid? Or 2000 psig (system pressure?)

The spiral wound gaskets are rated to 2000 psig - but that is very definitely inside pressure to outside pressure (usually zero!). Each spiral wound "V" is nested from inside to outside so the "V" resists leakage from inside to out. The other way, across a small psid, a standard spiral-wound will help by eliminating bad fitups and slight waves in the two faces. maybe some small scars and imperfections as the gasket faces settle in and get compressed.

But NOT 2000 psid.

 

[JJPellin]

The actual differential pressure is probably closer to 1800 psi. Pressure on the OD is about 1850 psig. The pressure on the ID is about 50 psig. I think this is a design error. When I take that back to the pump manufacturer, I would like to have some references to back up my position. I could not find anything that specifically excluded this.

Johnny Pellin

 

[sjrfc2]

I was working on a specific problem with spiral wound gaskets years ago where the pressure was from the outside in and I don't remember the chevron direction being discussed in any of the work I went through. What seemed like an issue for my problem was that the gasket was in a joint that was pressure loaded (versus unloaded like a flange joint). It was in a boiler superheater handhole joint and boiler pressure acted to compress the gasket further while it was in service. To compound this, the superheater temp was approx. 950 deg F and this would cause the yield of the gasket windings material (CRES) to drop allowing additional compression for the same load. I'm pretty sure I was able to get some test reports from WRC that showed that effect and the effect of creep from additional cyclic loading.

The combination of temp and cyclic pressure loading during operation resulted in the joint losing load over time and leaking. The location was accessible during operation and didn't lend itself to maintenance tightening very easily. The solution we tried was to use Belleville springs to try to maintain load as the gasket went through creep.

If you have pressure on the outside I'd recommend you analyze the joint to see if pressure is compressing the gasket further during operation (so that it may be loose when off-line). I'm not sure if Bellevilles would help in every application, there is an active thread in this forum right now that some people have relayed their experience.

I would also evaluate your joint: Is there a gage ring to limit compression? Is the gasket in a groove that will be limiting regarding compression? I would also check what you should expect from your bolt load vs your gasket rating and whether the bolt load (and pressure load if pressure loaded) is a good fit for the gasket. Gasket manufacturer's change the winding density to accommodate different bolt loads dependent on gasket rating. I've seen specs that have load vs deflection requirements for spiral wound gaskets so that they're matched to bolt load.

 

[sjrfc2]

Sorry, I meant to say the joint wasn't accessible during operation.

 

[gasketguru]

No - spiral wound gaskets are not particularly pressure energised e.g. like a hydraulic cylinder chevron set etc. - they can seal external pressure and vacuum as well as internal pressure. The V profile allows them to be wound and be stable of course as well as fold upon compression. As they compress, the filler reacts to the steel strip, so you are tightening the coil into itself (think like pulling a tape measure tight...) The windings do set up a radial internal stress, which is why some of the older ASME size gaskets with narrow inner rings suffered inner ring buckling.

Is your pump gasket simply a sealing element in a tongue & groove recess or does it have an inner ring etc. - If it was not fully compressed initially so the flanges come metal-metal, then additional external hydrostatic forces could have added to compression and thereby loosened the bolts perhaps? Worth also checking gasket to recess clearances - when compressed the recess can help support the outside of the windings and increase density. You need to analyse the joint as a whole of course.

 

[JJPellin]

The gasket in my application is trapped in a tongue and groove recess with a set compression. The bolting that locks the two faces in position is the same bolting that retains the pump head. The bolting is massive, as a result. I need to reevaluate the gasket compression. My application is actually more complex than I have described. My first interest was to confirm the design based on external pressure. I will follow up on Monday with additional design concerns.

Thank you for the very helpful replies.

Johnny Pellin

 

[JJPellin]

The second design concern I have is related to the fact that they actually have a stack of gaskets in this joint. In fact, they have two stacks of gaskets. These stacks consist of 7 spiral wound gaskets separated by thin spacer plates. These stacks are supposed to provide a preload to hold the inner casing of the pump in against another spiral wound gasket at the opposite end. I have attached a partial drawing showing the arrangement. When I approached the pump manufacturer about our problem, they initially claimed that these gaskets did not seal any differential pressure. I had to make up this sketch to show the fact that they are actually sealing full pump discharge pressure from leaking back to the chamber that leads to the balance lines. So, the two stacks of gaskets and spacers actually have to seal the full differential pressure of the pump.

So, in addition to my concern about the external pressure, I am concerned about the fact that 7 gaskets with 6 spacers and two sealing faces create 14 potential leak paths in parallel. The compression on this stack is intended to properly compress the gaskets, but I still need to take all the measurements necessary to prove that the compression is adequate for this gasket.

Any other suggestions or comments will be welcome.


Johnny Pellin

 

[MikeHalloran]

I can't say I've ever seen Flexitallics stacked like that. I assume that the reason for doing it is lost in the mists of antiquity.

The part that troubles me, beyond the expense and extra leak paths, is physical stability of the stacks, given the apparent thin-ness of the inter-stack spacers. Do they come out nice and planar, or wrinkled and warped?



Mike Halloran
Pembroke Pines, FL, USA

 

[JJPellin]

The spacer plates are thin sheet metal, probably 1/16" (0.062") thick. They are stamped plate and tend to dish and distort. I believe they are trying to use these gaskets for three purposes. First, they act to seal discharge pressure from leaking back to suction. Second, they act like a spring to load the inner casing into the back of the outer casing. Third, they accommodate some amount of differential thermal expansion between the inner and outer casing as the pump is heated up and cooled down.

I would very much like to eliminate these and replace them with an o-ring arrangement that can fulfill the first and last of these purposes. It will take some work to come up with something to satisfy the second purpose. I may end up leaving one stack of these gaskets to act like spring and replace the second stack with an appropriately designed spacer ring with o-rings. The process temperature never exceeds 375 °F.


Johnny Pellin

 

[gasketguru]

You are right - the stack of gaskets is basically a crude compensator, to take up build tolerances and thermal movement - usually at least 6-10 gaskets. (We used to do these many years ago for companies such as Mather & Platt, back when the gaskets were still asbestos!...) Whilst you might get some minor leakage through the stack and it represents inefficiency, the volume flow rate should be minor compared to the amount of fluid being pumped. Of course it is the main outer gaskets that are more important to avoid external leakage.

If you are talking to your local gasket manufacturer, then maybe worth asking if the stainless winding strip they use is available as "half-hard" material (as an alternative to fully annealed), as this should add a fraction of extra recovery to the gasket behaviour, or maybe try Inconel windings etc.

 

[JJPellin]

gasketguru,

Thank you for the suggestion. If I am unable to eliminate this stacked-gasket arrangement, I will certainly consider material changes to make it more effective. For these particular pumps, I have found three sources of inefficiency from internal recirculation. I believe that this gasket configuration is a significant source, but probably not the primary source of our losses. But, I plan to attack all three.

Johnny Pellin