MultipleSupport excitation

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

Moderators: silvia, selimgunay, Moderators

Post Reply
Jayram

MultipleSupport excitation

Post by Jayram » Mon Oct 04, 2004 2:40 pm

Hi,

I'm trying to run an analysis using MultipleSupport pattern. When I define the excitation in the form of a displacement, I get expected results. However, when I do the same analyis using the corresponding acceleration excitation, I get some garbage as output. (If I look at the output file, it has numbers of the order 10^-16 or so.)

Incidentally, the excitation I'm using is harmonic. Thus, acceleration = (w^2)*displacement. However, when I look at the output file, this is not true. Could this be due to numerical errors in computing acceleration from displacement?

Thank you
Jayram

fmk
Site Admin
Posts: 5883
Joined: Fri Jun 11, 2004 2:33 pm
Location: UC Berkeley
Contact:

Post by fmk » Tue Oct 05, 2004 11:03 am

multi-support excitation is implemented using sp constraints to specify the nodal displacements at each time step .. if you provide the acceleration record, numerical integration using the trapezoidal rule at present is performed to obtain the displacement record .. for those who don't know this, THIS IS NOT GOOD .. for this reason you should provide the displacement record when doing multi-support (the vel and accel can also be provided in addition to the displ record) .. the manual needs to be updated to state this, sorry for the confusion but i had thought it had already been taken care off.

Jayram

Post by Jayram » Tue Oct 05, 2004 1:01 pm

Hi,

I'm using the output from another program as input in Opensees. The other program unfortunately gives only the acceleration record. I'm unable to use UniformExcitation Pattern because different points in my structure are subjected to different acceleration histories.

You say that integrating the acceleration record to get displacement is not good. Are you specifically refering to the Trapezoidal rule, in which case I can use a better integration algorithm to get the displacement history and specify that in my Opensees analysis.

Thank you
Jayram

fmk
Site Admin
Posts: 5883
Joined: Fri Jun 11, 2004 2:33 pm
Location: UC Berkeley
Contact:

Post by fmk » Tue Oct 05, 2004 2:47 pm

yes i am referring to the trapezoidal integration strategy, which is all i have had time to implement and thus is the default, as being bad.

Post Reply