Ellipsoidal head with ts/L < 0.002

[TylerM]

I’m evaluating the design of a pressure vessel and it is constructed with an ellipsoidal head that is only 1/16 in thick (in the fully corroded condition). I’ve calculated the value for L (Inside spherical radius) and have come up with 40.63. The ratio of these two numbers (0.0625/40.63 = 0.0015) is less than 0.002 and because of that I have to justify the thickness using Mandatory Appendix 1-4(f). Problem is, when I work through Mandatory Appendix 1-4(f) I get to the conclusion that the head is too thin to hold my design pressure and I need to increase the thickness of the head.

Ignoring the requirement of ts/L > 0.002 and using the equation from UG-32(d) or Mandatory Appendix 1-4(c) the required thickness of the head is only around 0.03 in. Can someone please explain to me why one part of the codebook is saying that the head is too thin while the other part is saying the head is adequately thick? Or help me understand why the cut off for the ts/L ratio is 0.002?

Thank you for all of your knowledge,
Tyler

 

[pdiculous963]

When an ellipsoidal or torispherical head gets thin there is a concern for buckling of the knuckle. The reason the Code book says two different things is because you are evaluating two different failure modes. Appendix 1-4(c)/UG-32(d) is an evaluation for plastic collapse. Appendix 1-4(f) is evaluating buckling.

 

[JStephen]

My guess- thicker heads fail by yielding, thinner heads fail by buckling in the knuckle area and that t ratio represents the transition between the two cases.

 

[TGS4]

Check your thickness using the calculations in Div. 2 (assume you're using CC 2695). Those calculations will differentiate between the various failure modes (typically knuckle buckling for those extremely thin heads).

 

[r6155]

Remember min. thickness in UG-16.

Regards
r6155