Internal conical part in COMPRESS

[zjliang]

Hello Tom,

A question about how to use COMPRESS:

I am trying to design a vertical vessel with an internal conical section at the bottom 2:1 SE head.

The reason why I don't use a conical head at the bottom is the half cone angle is too big, almost 60 deg. So, in order to satisfy the process request, I am planning to design an internal conical secion.

As an internal part, the cone is not a pressure retaining part and located close to the bottom 2:1 SE head (estimated conical depth is close to the depth of elliptical head).

Note: in COMPRESS, i found that the cone is always treated as a transition pressure retaining part. How can I model an internal cone in COMPRESS? If there is no way to do it, I guess i have to add some additional weight on to the head. Is that right? Some other advice?

Thanks a lot in advance,

 

[TomBarsh]

Good Afternoon zjliang,

COMPRESS provides various types of pressure retaining components, including formed and flat heads, cylinders, and transitions (cones). Each type of component is designed per the relevant rules in ASME Code.

COMPRESS provides several options to specify non-pressure retainng "internal" components of various types, such as trays, packed beds, linings, insulation, etc (see the "Attach" menu). These types of components are not "designed" by COMPRESS; no strength calculations, etc, are produced for them. These components are provided only to provide a method of specifying enough information about the attachment so that COMPRESS can determine the weight of the attachment. The weight of the attachment is then applied to the vessel as a distributed load around the shell perimeter (in some cases, the eccentric nature of the load is considered, such as for platforms that do not extend totally around the vessel).

COMPRESS also provides options to directly specify vertical and horizontal loads that act on the vessel. These loads could be specified in lieu of specifying the details of the attachment components discussed above. These loads can also be used to specify a force that acts on the vessel that cannot otherwise be modeled as an attachment component. See the "Loads" menu. Note that "lateral forces" are treated as a pure force, with no associated mass; "vertical loads" are treated as a force arising from the product of mass and gravity, the mass portion is considered when determining the vessel natural frequency and vessel response for seismic conditions.

For your case you can determine the weight of the cone and manually enter this as a vertical load acting at the elevation at which the internal cone will be installed. COMPRESS will consider the weight of this component in the vessel calculations but will not "design" the part itself.





Tom Barsh
Codeware Technical Support

 

[zjliang]

Thanks a lot, Tom

I did exactly as what you described here, manually input a vertical force to consider the effect of the internal cone section.

BTW, I am still feel curious about the cone w/ 30+ deg half apex angle, w/o knuckle. I found that COMPRESS refuses to do the code calculation if there is no knuckle specified.

So the special analysis has to be done by hand, right?

Thanks again

 

[TomBarsh]

Good Afternoon zjliang,

What Code calculation are you expecting to see for your cone that are not reported?

COMPRESS performs the calculations for transitions with or without knuckles per rules of UG-32 and UG-33. Calculations are performed for the cone-cylinder junction as per Appendices 1-5 and 1-8. Note that the knuckle-to-cylinder juncture is exempt from reinforcing area requirements, but if the juncture is a line of support for external pressure then the moment of inertia requirements of Appendix 1-8 must also be met.

Per both Appendices 1-5 and 1-8, if the included angle of the cone exceeds 60° (ie: alpha > 30°) then additional analysis of the discontinuity stresses at the juncture is required. COMPRESS performs these under the heading of the "Bednar" calculations, referring to the text by Henry Bednar, whether or not there is a knuckle (or a flare at small end).

If you continue to have questions on this issue, please send your example file to me at Tom@codeware.com.

Tom Barsh
Codeware Technical Support