Continuous Concrete Beams

[Stazz]

I noticed this in 3d:

Recently I was modeling continuous grade beams by spanning 5 individual beams node to node on simple supports as individual physical members between each joint. The ends were modeled as fixed and the beams were in a straight line so I would suspect that the end moments would match the end moments of the adjacent beam but I noticed that these moments were a lot different. When I changed the model so that the beam was one continuous physical member over the intermediate supports, then I got the theoretical results I suspected where the full end moment transfered to the next beam.

After some dabeling, I found that I could get the 1st model to produce the results I wanted by releasing the axial load in each members. Is there some FE analysis theory behind this of why this causes the model to work?

 

[JoshPlumSE]

Stazz -

There is not anything that I know of that would explain the results that you were seeing.... Other than to say that I'm guessing there is a modeling error in the first model.

Quick tips on looking for the modeling error:
1) Make sure you haven't accidentally defined a rigid diaphragm.
2) Run a model merge.
3) Make sure you are not using the Top of Member offset.

Josh

 

[Stazz]

Yes, it looks like the Top of Member offset was on which caused this problem.

I wonder why that would create different moments at continuous ends along with a moment at the discontinuous end supported by a pin.

 

[Stazz]

Oh, ok, I see, so Risa physically offsets the member by attatching a rigid link to the end of the member that extends from its centroid to it's top fiber. So the end of the member isn't exactly free and the moments won't theoretically appear to jump over to the adjacent beams. It looks like these differential moments are created from the axial loads in the beams which are creating moment from the eccentricity of the rigid link. Thats why releasing the axial loads in the beams removed that problem.

But why are axial loads introduced into these beams when I have no external forces defined that would induce this load, I only have a single gravity load on the beam. The only explanation I could think of would be from catinary action but the supports aren't rollers, they are pins which should prohibit such action. I also don't have any gravity factor in the longitudinal direction.

 

[JoshPlumSE]

Yes, that is it... the program creates a rigid link from the node down to the centroid of the beam.

Then any boundary condition are defined to the original node (which is at the top of the beam).

If you make the model using rigid links (rather than TOM offsets) then you will see exactly what is happening and why.

Normally, the TOM offset creates more confusion than it is worth. I don't generally recommend using it unless you really know what you are doing. In fact, that is the one feature that I would like to have removed from the program. It's just so rare that folks use it in a way that provides them the type of results they want to see.