Pinned Base vs Fixed Base - Steel Frame

[Jacst3]

I am designing multiple steel framed structures to support equipment and piping. I am analyzing the structures in STAAD. There is debate on how to model the steel column bases: Fixed or Pinned. A truly pinned base usually results in excessive and unrealistic deflections. I have found a good compromise is to use fixed bases in the model, with partial moment releases at the bases for Mx, My and Mz.
A 1.0 factor equals no moment restraint. I've been typically using a factor of 0.9 or 0.95, meaning that the base is almost a pin. This seems to give reasonable results. The frames still require some bracing or moment connections to control deflections. Any comments or suggestions out there?

Thanks,

Jacst3

 

[delagina]

I assume you work in Petrochemical. If it's equipment structure then you can brace it on both longitudinal and transverse so assuming as pinned base is ok.

The rule of thumb in the previous company I worked, if bolt is outside it's assumed fixed. And I didnt work for some small company, it's one of the biggest EPC in the world.

I'm pretty sure some will come here and talk about base plate bending, foundation rotation, A bolt elongation, etc..

 

[Jacst3]

Delagina,

Thanks for your input. When you say "if the bolt is outside.." do you mean if the anchor bolt pattern at the column base is outside the footprint of the column shape then it is considered fixed?

Jacst3

 

[delagina]

yes, that was the assumption. piperacks, equipment structure, etc were all fixed. just 4 bolts outside the WF.

if it was a wrong assumption then 30+ str. engineers and some more overseas just made a wrong assumption on a multi-billion dollar project.

 

[rb1957]

but wouldn't you "only" get unreasonable deflections if the frame had a flexability in it, eg you're relying on the base fixity/friction to brace the frame ? i'd've thought that pretty much any pattern of fasteners would give you an essentially fixed joint, but assuming essentially pinned is conservative ...

 

[paddingtongreen]

I never used fully fixed because I couldn't provide an infinitely stiff foundation. After some experimental calculations, I started using, as a general rule, 50% for stiffness but 100% for strength. This was in the refinery design business, but before computers were available. I can't say how it worked in practice, I didn't have any failures, but i doubt they ever felt the full design loads. You should be able to model the foundation and get a feel for it's stiffness to feed back into your structure model.

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[csd72]

For portal frames we used to use totally pinned for strength but use superposition to calculate what fixity was required to keep deflection down to below the allowable limit.

Base plates are inherently flexible due to the oversized holes and the fact that the anchor bolts cannot be pretensioned. I would never use 100% stiffness for these in a portal frame.

In a direct cantilever then you have no choice and have to treat it as fixed, though you need to realise that there will be more deflection than calculated.

 

[delagina]

paddington,

staad has KMX KMZ KMY with units of

kip-ft/deg


so what's a good value here to assume for partial fixity

 

[rscassar]

If it has been designed as fixed (i.e. baseplate, anchor bolts and foundation) then I would use a fixity factor of 0.9 for serviceability load case and 1.0 for ultimate. It reality, there is always going to be some rotation at the baseplate. As soon as a little bit or rotation occurs at the base there is a big redistribution of moments.

This has been discussed before, I will dig around and post the thread numbers when I find them.

 

[apsix]

There's some discussion here on the British Standard BS5950 approach:

http://www.eng-tips.com/viewthread.cfm?qid=260012&page=3

 

[csd72]

kikflip,

I completely disagree with your logic here, if the plate is not fully fixed at serviceability state then it is definately not fully fixed at ultimate.

The items that compromise fixity such as footing rotation, anchor bolt elongation and plate bending with all be more pronounced at ultimate state and therefore your factors should be the other way around if anything.

 

[rscassar]

If you are trying to limit drift at serviceability limit state it would be an appropriate assumption to use 90% base fixity to model the additional drift that is caused by the rotation at the baseplate.

Likewise if you are designing the baseplate at ultimate limit state I would want the baseplate and anchors to transfer the moment that arises from full fixity. Not a reduced moment because some of the base moment has been redistributed elsewhere. I would multiply the reactions at the baseplate by 1.05 for baseplate design.

 

[strurctualengineer]

Hi all,

Can somebody please explain why BS5950 (UK steel code)
appears to only allow partial fixity in column bases upto 20 percent?

If i understand correctly, a partially fixed base does not induce a moment as a percentage of the fixed base moment; but rather is determined by the column stiffness itself.
Therefore can there be a situation where say taking a 70 percent partial base fixity would induce a larger moment than a fully fixed base?

But I also was of the understanding from moment distribution analysis that a fixed base had 'infinite' stiffness, in that case a partially fixed base moment could not be greater than a fully fixed base moment.

If anyone can offer any guidance into these matters it would be much appreciated.