Vessel legs attached to head on PV

[NovaStark]

For a PV, I've read that legs are generally attached to the shell, however there is little explanation as to why.

For anyone with experience, can you tell me your thoughts as to the design considerations for this leg design ?


Also if you have experience with PV Elite, does the software support this ? As I've attempted to try to find a way using the "leg" module however there isn't any distance from the node that I can enter that allows the legs to be in contact with the head.

 

[dhengr]

NovaStark:
When you apply the leg/column loads to the side of the tank, the shell, you are distributing the primary loading, an axial load, into the shell in shear. This is easy to do and analyze, and the length of the connection can be extended up the tank to lower the shear flow, which goes directly into the shell plate thickness. Also, you can accommodate some leg bending action caused by lateral vessel loading, over that longer lever arm. When you apply the same leg loadings to the head of the vessel, essentially perpendicular to the plane of the plate (or almost perp.), you are applying the leg loads as a punching compressive loading on the head, and as a plate bending load over the diameter (or size) of the leg and into the head plate. This latter condition, while not impossible, is much more difficult to rationalize. It quickly becomes a very complex plate bending problem, with several different additive plate bending loadings and punching shear into a plate which has complex geometry and boundary conditions. Not a good or easy design problem.

 

[GheorgheCristian]

Have a look at EN 13445-3 chapter 16.11.1 . Also a software that does that kind of calculation is Bentley Vessel. PVElite does not have the ability to model or to calculate legs on heads ( i tried and it just extends the legs throu the head to where they were supposed to be if they were conected to the shell)

 

[r6155]

See recent post

http://www.eng-tips.com/viewthread.cfm?qid=416255

Regards
r6155

 

[TomBarsh]

Buckling of the bottom head due to the concentrated loads at the legs is also a non-trivial problem.

 

[NovaStark]

Ah so then it's more for ease of design/analysis and possibly installation rather than legs on a head are more likely to fail.

 

[SnTMan]

NovaStark, yes, and a very useful concept to keep in mind. If you can't easily design / analyze it, maybe you shouldn't do it absent a really compelling reason.

Regards,

Mike


The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[NovaStark]

Thanks for the insight all!

 

[dhengr]

NovaStark:
Another way of saying what SnTMan said above is..., Clean, simple, straight forward, direct load path design is, more often than not, the best design. It often takes a bit more engineering time at first to search out a good clean design, but it generally saves money in the long run, in terms of fabrication, usage and maintenance. You can design a real Rube Goldberg master piece, and some people will say oh! what an amazing and complicated design, you’re a genius. But, this is often just layers of complex crap with the latter layers needed to cover up or resolve an earlier layer that you should have resolved and cleaned up in the first place, and just shows their lack of experience and knowledge of real good engineering.