Design of spherically dished covers (bolted heads)

[daveb9]

Design of spherically dished covers (bolted heads)

ASME IIIV Appendix 1 1-6

I am trying to optimise a design of a bolted head similar to FIG 1-6 (D) with regards to the thickness of the flange T.

My design differs from the figure (D) by the fact that the seal is carried within the flange in a groove, this groove would typically be 0.375 deep, and the bolting is achieved by lugs welded to the outside diameter of the ring.

Up to now our standard calculation technique would be to calculate the centroid using the flange area less the seal groove, then the calculate the thickness required this thickness is then said to be the minimum distance from the bottom of the seal groove to the top of the flange.

My question is does anybody know of a acceptable calculation which would allow me to consider the whole thickness of the flange not just the material above the seal as the flange thickness?

Thanks

Dave

 

[GenB]

I have never seen such a large groove on the holding/reinforcing ring, it if was just 1/8th it may not even be noticed in the calcs but you are talking about 3/8th of an inch and that's a lot, you may even account for half of that or 3/16th (or 10% of the total ring thickness)and reinforce the reminder and that's depending of the total thickness of the ring.
I will research further and post.
genb

 

[prex]

In my opinion you could take an effective thickness of
t_e=2A_F/(A-B) where A_F is the area of flange cross section (and of course taking away also any corrosion allowance from that).
This procedure is adopted by App.G of EN13445, that contains a quite complex method (50 pages) for calculating ring flanges, that is an evolution of App.2 method. In this method however the flange is still considered as a circular ring resisting by bending around an axis perpendicular to the axis of the ring.
If you can adopt that standard, I'm sure it would help you in optimizing your project (but personally never did such a calculation).
If you stay with Div.VIII I think you can still justify the above formula: in fact App.2 calculation relies onto the resistance modulus of the flange around a centroidal axis perpendicular to flange axis, and this one is proportional to the square of flange thickness: hence that way of calculating the effective flange thickness is conservative.
Unfortunately cannot propose a bibliographic reference for such a reasoning under ASME VIII coverage.

prex

http://www.xcalcs.com

Online tools for structural design

 

[daveb9]

Thank you Prex and GenB

The current route I am taking is to model the flange in a FEA package. I have calculated thicknesses from ASME of a standard flange with no groove, and have FEA results for the ASME design. I am now developing the shape of my deep groove design keeping the deflection and stress levels below that of the original ASME design.

Unfortunately the section of the ring is no longer constant, the OD is not round and the ID can only be estimated therefore the EN 13445-3 calculation G.5-7 may not apply

As a alternative I intend to proof test with strain gauges to ASME UG101.

Would you consider this to be a sound proposal.

Best Regards

David

Dave