ETABS - mass participation ratios

[dadomago75]

Hi all,
I have a complex model with a weird shape that gives me problems.
I managed to run the Modal analysis using Eigenvalues algorithm but the mass participation ratios comprise a big component of RX RY.
According to ETABS analysis reference manual UX UY UZ RX RY RZ are about the center of mass whatever that means.
I am not expecting RX or RY.
So what I did is to model a 3 storey structure regular in plan and elevation doubly symmetric with a wall box centered to the slab.
5 kPa UDL over the slab.
I setup 3 models with the following results:

Columns only

Columns and walls

Walls only

The columns only is obviously too flexible and all the mass is activated by the first two modes that are translation.
The other two are pretty similar as I would expect. They both have in common the first mode as torsional.
This worries me because the structure is doubly symmetrical and the center of mass coincides with the center of stiffness. I therefore not expect any torsional mode.

Has anyone came across this problem and investigated it?
Thanks
Davide

 

[Agent666]

The only way you are not going to have a torsional mode is if you restrain that degree of freedom. Simply aligning the centre of stiffness and mass does not restrain/eliminate torsional modes. Look up the mass moment of inertia, a 3D structure is always going to have a lateral amd torsional response.

Try sorting the column stiffness, the period is off the scale for those modes. Either something is unrestrained/flapping around, or the model simply is not solving due to being ill conditioned (check the warnings during analysis),or some of your properties are an order of magnitude out.

Locating your lateral force resisting system further from the centre of mass, or increasing their stiffness increases the torsional stiffness may help to ensure first modes are translational.