Cone to shell junction stresses

[MikeG7]

Hi I was wondering if anyone can shed some light on the concept of the stresses at the intersection of the shell and cone? My question stems from an observation that flatter the cone end, the greater the discontinuity (and stress):
when the half apex angle is less than 30 degrees, we don't check it in ASME VIII Div 1 calculations.
when it is more than 30 degrees, methods like 8.6 in Bednar are available for analysis of the stresses,
...when it gets to a flat cover, once again it is not considered unless it is by Appendix 14 for a cover with a large opening.
I would expect a rotation of the intersection of the flat cover and the shell intersection, more than with a cone so I also would expect higher stress on the shell at the flat cover end.
Just looking for some basic guidance for dealing with shell discontinuity stress when the cover is flat.

 

[marty007]

I ran a couple of FEA's out of curiosity a while back for flat heads attached to shells (so the following is going off memory).

As you are thinking, the peak stress in the model occurred in the shell right at the shell/head junction. That said, as long as you follow the rules in ASME VIII-1 for the head thickness, then the stress in the shell was reasonable. The head itself appeared to be much thicker than required, but this additional thickness was acting as a stiffener for the shell/head junction.

 

[MikeG7]

marty007
without doubt, that is where the discontinuity stresses are likely to be at a maximum. I don't think that there are any checks in UG-34 (at least none I can spot) that deal with the stress in the vessel.
I wonder why that is?

 

[marty007]

I think you missed my point.

Although nothing in UG-34 specifically discusses the discontinuity stresses, the formulas for the thickness of the heads inherently addresses this junction. In the formulas for determining the required head thickness, the value 'C' will add additional thickness to the head to reduce the shell/head junction stresses.

Think of it this way, in a cone/shell junction, one way to reinforce the junction is to add an external stiffening ring. With a flat head, it's normally a lot easier and cheaper to provide that additional reinforcement by just thickening the head a little extra (i.e. internal stiffening ring).

 

[MikeG7]

I've never had any problems in applying the rules of UG-34 as far as the shell is concerned, but it would be nice to know HOW it is accounted for without just knowing that the code is "taking care of it"? Put another way, it the head were formed with a very subtle taper say 1:50, it would no longer be a flat cover (can't apply UG-34) and one would have to quantify the junction stresses.