2d Dynamic non linear RC fiber model - Period too long!

Forum for OpenSees users to post questions, comments, etc. on the use of the OpenSees interpreter, OpenSees.exe

Moderators: silvia, selimgunay, Moderators

Post Reply
ptr
Posts: 5
Joined: Sat Sep 23, 2017 1:52 am
Location: università degli studi di padova

2d Dynamic non linear RC fiber model - Period too long!

Post by ptr » Mon Apr 09, 2018 1:19 pm

Hi everybody!
I am a senior university student working on my last project.
I tried to run non linear dynamic analises for a 2D RC fiber frame, based on some examples founded on web.
After working a lot trying to make the analyses work, I found a problem that I really can't solve! Long story short, my frame, under impulsive acceleration (i.e. a ground motion recorder with a single value not equal to zero), starts to oscillate with a huge foundamental period: more than 14 seconds!
Obviously, this problem happens even when I use natural ground motion recorder.
Masses and materials' parameters seems right after checking them more than once.
Did someone of you have the same problem or know which can be the causes?

Thank you!

BIal
Posts: 14
Joined: Wed Jun 21, 2017 7:43 am

Re: 2d Dynamic non linear RC fiber model - Period too long!

Post by BIal » Mon Apr 09, 2018 1:54 pm

Before running the dynamic analysis, I would suggest you to simply try an eigenvalue analysis before gravity loading and after gravity loading (if you are applying gravity loads). See if you find anything strange by comparing the results.

ptr
Posts: 5
Joined: Sat Sep 23, 2017 1:52 am
Location: università degli studi di padova

Re: 2d Dynamic non linear RC fiber model - Period too long!

Post by ptr » Tue Apr 10, 2018 11:38 am

Thank you so much for your answer!
I did what you suggested me, but it seems that the problem has nothing to do with the application of gravitational loads! Indeed, there are no differences between the results of the eigenvalue analysis before gravity loading and the results after it. I think that the problem is very common, but I really can’t find any solution!
Do you have any other suggestions?
I now attach the values of the eigenvalue analysis that as you can see, are equal before and after the gravity loading and with huge periods:

periods are 12.777223602266577 3.279120727316422 1.895721017992266
eigenvector 1: 1.0000619265446669 1.0000619265446669 1.0
eigenvector 2: 1.0013569457670897 1.00135694576709 1.0
eigenvector 3: 1.0062549954782252 1.0062549954782252 1.0
Application of gravity load
Gravitational analysis
Gravitational analysis done
periods are 12.777223602266577 3.279120727316422 1.895721017992266
eigenvector 1: 1.0000619258737504 1.0000619258737495 1.0
eigenvector 2: 1.001356938153136 1.0013569381531349 1.0
eigenvector 3: 1.0062549741294982 1.0062549741294973 1.0

Thank you for your time!

selimgunay
Posts: 883
Joined: Mon Sep 09, 2013 8:50 pm
Location: University of California, Berkeley

Re: 2d Dynamic non linear RC fiber model - Period too long!

Post by selimgunay » Tue Apr 10, 2018 11:49 am

Result of the eigenvalue analysis is not the period. It is w^2. You need to compute the period T as 2*pi/w.

ptr
Posts: 5
Joined: Sat Sep 23, 2017 1:52 am
Location: università degli studi di padova

Re: 2d Dynamic non linear RC fiber model - Period too long!

Post by ptr » Tue Apr 10, 2018 12:07 pm

selimgunay wrote:
> Result of the eigenvalue analysis is not the period. It is w^2. You need to
> compute the period T as 2*pi/w.

Yes, I know it!
Indeed the period has been calculated in this way and unfortunately it is confirmed by the displacement's plot!

selimgunay
Posts: 883
Joined: Mon Sep 09, 2013 8:50 pm
Location: University of California, Berkeley

Re: 2d Dynamic non linear RC fiber model - Period too long!

Post by selimgunay » Tue Apr 10, 2018 2:17 pm

Please check your units. If they are fine, please post your script here

Jhno
Posts: 214
Joined: Sat May 05, 2012 2:55 pm
Location: Laval

Re: 2d Dynamic non linear RC fiber model - Period too long!

Post by Jhno » Tue Apr 10, 2018 2:50 pm

The solution to the eigen problem involes the M and K matrix. In my understading (which might be wrong of course) unless your load bring a modification to the rigidity matrix (yielding) there is no reason that application of a load change the outcome of the analysis.

ptr
Posts: 5
Joined: Sat Sep 23, 2017 1:52 am
Location: università degli studi di padova

Re: 2d Dynamic non linear RC fiber model - Period too long!

Post by ptr » Tue Apr 10, 2018 10:13 pm

Jhno wrote:
> The solution to the eigen problem involes the M and K matrix. In my
> understading (which might be wrong of course) unless your load bring a
> modification to the rigidity matrix (yielding) there is no reason that
> application of a load change the outcome of the analysis.

I think so too.
Can someone tell me what influence so much the calculation of the stiffness except for the parameters of the materials and the geometry?

Jhno
Posts: 214
Joined: Sat May 05, 2012 2:55 pm
Location: Laval

Re: 2d Dynamic non linear RC fiber model - Period too long!

Post by Jhno » Wed Apr 11, 2018 6:03 am

I would check three thing.
1-ensure you have good fix condition and draw the frame to make sure node coord. and element are placed where you intended them to be (you can use a post-treat like GiD to show the model).
2-Check for the mass (make sure they aren't too large?)
3-Check the definition of your materials (did you provide consistant units?) and check your fiber section (check if you defined them well. Here is an example for fiber construction : https://searchcode.com/codesearch/view/43081462/). Also check the rigid joint connection (usually use a elastic element with high stiffness to connect beam and column to ensure the rigidity of this connection... how are these elements defined in your model? If it's too weak, the whole model will be weak and hence have long periods).
Just think about w=sqrt(K/M). Easy to check if M is correct, but alot of thing can affect K.
What constraint handler do you use btw? Transformation?
Right now, your model either have too high mass or too small K (in proportion to each other).

GL

ptr
Posts: 5
Joined: Sat Sep 23, 2017 1:52 am
Location: università degli studi di padova

Re: 2d Dynamic non linear RC fiber model - Period too long!

Post by ptr » Sun Apr 15, 2018 3:33 am

The problem seems to be solved!

There were some problems about the units and plus I didn't use any kind of rigid joint connection. Based on the examples founded on the web, I didn't think that it was necessary.
Is it really always necessary to use them?
Thank you all for your suggestions!

fmk
Site Admin
Posts: 5866
Joined: Fri Jun 11, 2004 2:33 pm
Location: UC Berkeley
Contact:

Re: 2d Dynamic non linear RC fiber model - Period too long!

Post by fmk » Mon Apr 16, 2018 9:44 pm

you do not need units at all .. the examples have them because the person who did them firmly believed in them .. if you keep your units constant in a file then drop them.
as for rigid offsets .. they may be needed, depends on the column and the connection, if very stiff then modeling the beam going ceneterline to centerline results in a structure more flexible than it really is.

Post Reply