Issue with Staad.Pro modal calulation

[atrizzy]

In Staad (and I may be mistaken about this) a full length beam, say L=3m should be identical to a beam with a node placed, say at the midpoint (such that L=1.5+1.5m), provided that at that node there are no releases in shear, moment or torsion.

However, I've noticed that mode shapes seem to consider deformation at the node regardless of the fact that that 'node' should be as uniform as a beam without such nodes.

Am I missing something? Any idea what I'm doing wrong?

Thanks very much.

 

[WARose]

However, I've noticed that mode shapes seem to consider deformation at the node regardless of the fact that that 'node' should be as uniform as a beam without such nodes.

No that isn't right. One of the most common mistakes I see in a big model (for a time-history analysis) is a lack of nodes along a member. The more flexible it is.....the more nodes you need. For a dynamic analysis, no nodes=no results. And that can prove to be deadly. While it is true that STAAD (and a lot of other programs) will interpolate results between nodes to (for example) give shear/moment diagrams/forces in a static analysis.....not so with a dynamic analysis. Everything has to be meshed (including beams).

 

[JoshPlumSE]

I agree with WARose's basic premise. If you want to get a beams mode shapes, you almost always have to add nodes along it's length. There are a couple of reasons for this:
a) If you want mass to be associated with the middle of the beam, then you generally have to have a node there. Because most programs will discretize the mass into lumped nodal mass.
b) Most programs do NOT automatically add rotational mass at the joints. I don't believe this is something that STAAD does (though I could be wrong). As such, even if there are rotational degrees of freedom at the ends of the beam that would capture a couple extra modes of vibration, those modes cannot be captured is there is not either translational mass (i.e. nodal mass) along the bam or rotational mass at the ends.

That being said, there are plenty of times where you do NOT want to add these nodes along the beam because you are more interested in the global response of the structure rather than the local response of an individual beam.

 

[WARose]

That being said, there are plenty of times where you do NOT want to add these nodes along the beam because you are more interested in the global response of the structure rather than the local response of an individual beam.

Right. For some type of seismic analysis....you really don't need it. But if someone is modeling a structure because we are talking a time-history analysis to see the response of a structure to a piece of machinery....yeah, you need it. I just fixed one structure where the (original) EOR didn't mesh the vertical braces in his time-history analysis. And guess what? The braces started flopping all over the place when the compressor on the 2nd floor was turned on.

 

[atrizzy]

Thanks folks, answered my question and then some.