Knife-Edge Flanges

[Eltron]

I tried this post over in the Pressure Vessel forum, but all I heard back was the wax building up in my ears. So I thought I'd give it a try over here.

I've got a question that I haven't been able to answer, so I thought I'd throw it out to the experts in the field. I'll preface my question by saying that I am not doing this, I just am unsure why it isn't done. More curious than anything else. So here's the question (hopefully it isn't too idiotic): why shouldn't flanges that use knife-edge seals (such as con-flat flanges) be used in pressure vessel designs?

I understand that CF Flanges are designed specifically for vacuum applications, and the manufacturer hasn't tested them for pressurized applications. From one manufacturer's website: "Positive pressures are inherently dangerous and failure comes without warning. In a vacuum system, failure depends on various strength properties of bolts, clamps, chamber walls, welds, feedthroughs, valves, viewports, etc. The only safe answer is that the internal absolute pressure cannot exceed the external absolute pressure." The real crux of my question is where will the CF flange fail? Is it the copper gasket? I'm hoping for suggestions other than "Because the manufacturer said so," or "Review the Boiler and Pressure Vessel Code."

I tried to contact the engineers at the manufacturer, but I just got a lot of CYA statements instead of actual physical reasons that this type of flange is a no-no.

Dan

http://www.eltronresearch.com

Dan's Blog

 

[handleman]

One guess: The pressure differentials are orders of magnitude difference. CF flanges are designed for very high purity, but very low pressure differentials. 14 psi is pretty low differential. Also, the vacuum tends to pull the knife edge into the seal rather than pushing it out. I doubt it would hold up well against internal pressure. Any "flex" in the flange would pull the knife edge out of the copper.

-handleman, CSWP (The new, easy test)

 

[rb1957]

i'd also question the tension capacity of a CSK head vs a plain head. there is also, of course, the question of fatigue loading.

the practice may be something prohibited by your code due to bad experience in the past. it could also be something that you can prove (by analysis and test), particularly if it is designed to be fail-safe (lose one fastener).

 

[handleman]

CSK head? Where does that come in?

-handleman, CSWP (The new, easy test)

 

[rb1957]

what else does "knife-edge" refer to ? i'm guessing it's something less in part 'cause of the query, in part 'cause the drwgs i saw on-line didn't show a CSK.

 

[handleman]

CONFLAT knife-edge seal is a type of flange seal used in high-vacuum and ultra-high-vacuum applications. It's a metal-to-metal seal that has high temperature resistance, almost no leakage, and low permeability/outgassing.

Here's a link to a PDF with a brief explanation:

http://www.caburn.co.uk/resources/downloads/pdfs/sec1.1.1.pdf

-handleman, CSWP (The new, easy test)

 

[rb1957]

thx ... the pix made it clear to me that these things won't work with +ve pressure, no?

 

[handleman]

That would be for someone other than me to say. These flanges are certainly not intended or designed to contain positive internal pressure. That said, they would, to a degree, contain a positive pressure. I'm not sure how well. These specific flanges are designed so that the stainless flanges don't actually contact one another. That is not particularly easy to see in the PDF linked, but it is stated in another catalog I have. Without flange-to-flange contact, I would think that you might not get enough preload on the bolts to keep a high positive internal pressure from separating the joint to enough degree to break the seal.

I guess my question to Dan would be, why would you want to use these flanges against internal pressure?

-handleman, CSWP (The new, easy test)

 

[hacksaw]

primary mode of failure is work hardening of the copper gasket, so you replace them. they are not reliable pressure seals


If you are going to see positive pressures you need a better design than trying to use CF fittings.

 

[Eltron]

handleman,

I can see how the CFs would be attractive in laboratory situations: high temperature rating, compactness and chemical resistance. My guess is that why they are avoided in industry mostly has to do with leakage as you mentioned above. I was just wondering what the other potential problems might be.

Dan

http://www.eltronresearch.com

Dan's Blog