Vacuum Vessel Flange Design 1

[rjw57]

I am involved in the design of vacuum vessels which operate from full vacuum to as high as 150 psig. We currently have a PE design our vessels which fall outside the "B&PV Code not required range". We have a vessel with a hinged front door so that the door can swing open for access to the interior (think horizontal autoclave vessel front door). Without getting into the particulars of the flange clamping method (assume bolting for the moment), our flange is sealed using a self energizing elastomeric precision o-ring nestled in a groove cut into the "door" flange (obviously, the seal is inside the bolt circle). There is full metal to metal contact along the faces of the mating flanges when the door is closed. My question is this - for code designs, our PE has considered this to be an Appendix 2 flange. It is my opinion that this should be construed to be an App Y flange design. Who is correct?

Next, I realize that App 2 can accommodate external pressure forces which considers prevention of buckling (or at least I believe it does). I don't see any such allowance in App Y calculations. Is there a way to consider a full vacuum when designing App Y flanges? If not, does this fall under the idea of "use your best engineering judgement" part of the code? If so, can anyone suggest a methodology which would be applicable. Hope this is all clear. Thanks for your help!

 

[prex]

Flanges with contact outside the bolt circle are explicitly excluded from the scope of App.2, so I think you are correct in pointing to App.Y.
Not sure to understand what you mean with buckling being addressed in App.2 and I think that App.Y may be easily adapted to external pressure by following the same lines as given by App.2. Note also that flange and bolt stresses are near to zero for a full contact flange under vacuum.

prex

http://www.xcalcs.com

: Online engineering calculations

http://www.megamag.it

: Magnetic brakes and launchers for fun rides

http://www.levitans.com

: Air bearing pads

 

[rjw57]

@prex

Thanks for the confirmation on App Y

In response to "Not sure to understand what you mean with buckling being addressed in App.2", I would assume that external pressure on any cylindrical object would subject this object to potential buckling since there would be compressive hoop stresses when the flange is externally pressurized. I was just saying that I think App 2 calc's would accommodate these forces (in external pressure calculations), however, these stresses may NOT be the ones that are the failure mode. It may have been better had I originally said nothing about buckling in the original post.

Here's the part I don't understand about your reply: you say "I think that App.Y may be easily adapted to external pressure by following the same lines as given by App.2." Can you please point specifically to what you are referring to in App 2 of the code with this sentence? I just am not clear as to where App 2 calcs can be applied to App Y calcs.

Finally, I understand the flange and bolt stresses being low under vacuum conditions. I guess I have to go back to the potential for buckling (even though I said I shouldn't have) when the flange set is externally pressurized. In some cases, we have chambers that are designed for as little as 3 psig internal pressure but also full vacuum. I cannot imagine that the design for 3 psig internal pressure would be adequate for full vacuum loadings. This is the problem I want to avoid. If I can figure out how to make sure the flange is safe under vacuum, I can turn around and say the chamber is safe for some given (yet to be determined) positive internal pressure as well.

Thanks for your reply!

 

[prex]

Concerning buckling, I think you should check your cylinder (or head) to external pressure, possibly using the flange as a stiffening ring, if it satisfies geometric constraints per code. That's all I think for flanges participating in buckling strength.
Concerning the other question, if you compare the formulae for M[sub]o[/sub] in 2-11 and 2-3 you'll see that under internal pressure the flange is considered as supported at the bolt circle, whilst under external pressure it is supported at the gasket.
A similar approach is used in App.Y (for internal pressure only), except that, if I recall correctly, the flange is supported at the OD of the contact zone.
In the same vein, under external pressure, an App.Y flange would be supported at the ID of the contact zone and also H[sub]D[/sub] and H[sub]T[/sub] would normally give counteracting (partly balancing) moments.
I'm not pretending that you should fully rework App.Y formulae for external pressure conditions, as, as an added difficulty, the basis for deriving those formulae is not given in the code. However you could devise a simplified approach to demonstrate that the stresses are negligible.
Concerning the flange strength when designed to external pressure only, it is true that a too thin flange could be derived from the above considerations. In those cases other criteria should be used to determine a minimum strength, such as lifting or clamping conditions. In my opinion the most straightforward way would be to define a minimum internal pressure to which to design your chambers: this wouldn't necessarily be stamped on the vessel, but only used in design.

prex

http://www.xcalcs.com

: Online engineering calculations

http://www.megamag.it

: Magnetic brakes and launchers for fun rides

http://www.levitans.com

: Air bearing pads

 

[rjw57]

@prex

Thanks for your input. You have given me a great deal of insight into my problem. I have decided to approach this first as an intermediate stiffening ring so that it is adequately sized for external pressure performance. Next, I will attempt an App Y design at the operating internal pressure. If the required cross section is not adequate to also serve as a stiffener, then I will increase my design pressure until I get adequate moment of inertia. The bolting will be designed for the operating pressure, not the design pressure (unless there isn't a great deal of difference). Maybe this is what you were trying to tell me in your last post. If so, I've got the message. If not, any advice on a potentially different approach?

Thanks again!