Out-of-Plane Element Movement in Eigenvalue Analysis

[Seismo123]

I have a few beams between shearwalls of an RC structure (coupling beams) of which have short free span smaller than their section height. Running Eigenvalue analysis, results in out-of-plane unusual motion of these beams while it is opposite with other couplings, ie. beams that do not have smaller free span than their section height. What would a possible approach to deal with this without changing sections' dimentions?

[huffte]

I cannot say I fully understand the situation Seismo123. I wonder if you could expand your description, provide screen shots, or send the model (thuff@tntech.edu) while I have a bit of free time if you wish?

[huffte]

I see that the issue may be eliminated in one of two ways Seismo123.

First, and most simply, since you are only doing a 2D analysis, go to the settings and set the mass directions to x, z, and ry only. Uncheck the other boxes. This indeed wipes out the spurious out-of-plane mode.

Secondly, you could experiment with the penalty factor weight in the constraints tab of the project settings and the number of Ritz vectors in the x, y, and z directions in the eigenvalue tab of the settings. See the user's manual for more discussion on these topics.

The first method is the most straightforward. But you and your supervisor will have to decide if the model is accurate or not. I suspect some very subtle difference in the rigid links you have employed in the 5th frame versus the 4th frame, where the spurious mode is absent. You could also dig into your model and try to identify such subtlety, if in fact it does exist.

Best of luck Seismo123.

[Seismo123]

Thank you Huff.
I chose the 1st way for the moment. Here comes fbd_Inv and fbd_Tol message on dynamic time history analysis now... eww
I have tried to increase the elements' convergence tolerance from the Elements tab (Tol.=1) and the global convergence tolerance values (Disp. Tol.=0.9 mm) from the Convergence Criteria tab. I mean how far I should go?

[seismosoft]

Seismo123, have you read the parts relevant to convergence issues in the manual? In v2018 there is a complete description of the divergence messages, as well as a step by step procedure to fix the convergence problems.
Seismosoft Support

[Seismo123]

Well... It has been a while since then, even I installed v2018, but I still do have the convergence issue which arises on dynamic time history analysis, but not on static time-history nor Eigenvalue Analysis. There should be something simple yet nerve-wracking though.

[seismosoft]

Eigenvalue analysis is an elastic type of analysis. There are no divergence issues in Eigenvalue analysis.
Regarding the convergence issues or not in dynamic or static analysis, these are obviously affected by the type and amplitude of loading. Please make sure that you follow all the steps proposed in the manual.
Seismosoft Support