Modelling Building floors (shell element)

jwieser · Post by **jwieser** » Wed Oct 14, 2009 1:34 pm

Hello,

I am modeling a three story building in three dimensions and am looking for the floor response due to triaxial excitation. I have a bare frame model that uses frame elements and floor described with a rigidDiaphragm command. I also have a model that explicitly models the floor diaphragm with shell elements. The building is a steel building so the shell elements are connected to the frame through the use of the equalDOF command so that the moment in the shells is released. Due to this connectivity I am unable to assign a rigidDiaphragm constraint to these nodes as there would be constraints in series.

My question is when I compare the primary modes of the two structures. The primary torsional mode of the building is lost when the shell elements are used. Is this due to this:
"Should the element be required to compute a mass matrix, a consistent translational element mass matrix is computed. Rotational-inertia terms are ignored." (this was taken from the mannual on shell elements)? even though I have assigned the mass to the nodes connecting the shells as opposed to using the rho factor in the shell section definition.

If anyone has any ideas I would be receptive.

Thanks

silvia · Post by **silvia** » Fri Oct 16, 2009 9:14 am

i have not done this type of modeling, yet, so i don't have much insight.
let us know how it goes, though.

jwieser · Post by **jwieser** » Fri Oct 16, 2009 9:37 am

I think that for some reason the shell elements are condensing out the I-Theta term of the mass matrix thus eliminating all of the torsional modes of the building. Is there an element like a shell element that I could use that would not do this?

Thank you