Hi, everyone!
I want to analyze an isolated structure with initial displacement on isolators. Should I add permanent load(displacement) on the isolators? If it is right, I wonder whether I should add the displacement on node1 or node2? I've tried this, but the response of the upper structure like storey drift was not obvious. Did I do something wrong?
Thanks a lot!
pre-displaced isolator
Re: pre-displaced isolator
Interesting problem Fionna Dong. I'm wondering how many isolators you have in your model and what type of analysis you are attempting.
If you are running a dynamic time history analysis, it seems that you could develop a simple sine wave loading of some arbitrary magnitude with several seconds of trailing zeroes. Apply this to the structure in a time history analysis. The reason for this is to get the residual displacement in the isolator when subjected to the specified load. Knowing this residual displacement, you could factor the input load either up or down to get the desired initial displacement in the isolator.
Then, finally, you could append your real time history loading to the modified sine wave (keeping the several seconds of zeroes between the two). Your structure will then be subjected to the real time history with an initial displacement equal to the residual displacement at the end of the zeroes.
It could be that you might get your method of applying a permanent displacement load to work properly as well, but like you, I have difficulty seeing how the mechanics of that work.
Best of luck Fionna Dong.
If you are running a dynamic time history analysis, it seems that you could develop a simple sine wave loading of some arbitrary magnitude with several seconds of trailing zeroes. Apply this to the structure in a time history analysis. The reason for this is to get the residual displacement in the isolator when subjected to the specified load. Knowing this residual displacement, you could factor the input load either up or down to get the desired initial displacement in the isolator.
Then, finally, you could append your real time history loading to the modified sine wave (keeping the several seconds of zeroes between the two). Your structure will then be subjected to the real time history with an initial displacement equal to the residual displacement at the end of the zeroes.
It could be that you might get your method of applying a permanent displacement load to work properly as well, but like you, I have difficulty seeing how the mechanics of that work.
Best of luck Fionna Dong.
Tim Huff
-
Fiona Dong
- Posts: 12
- Joined: 01 Dec 2013, 00:37
Re: pre-displaced isolator
Hi,huffte. Thank you for your help firstly.
I have 74 isolators in my model, it is a super-long isolated structure with irregular plane. I want to do dynamic history analysis.
I still wonder that if I develop a simple sine wave, should I add this real time history on node1 or node2? Since all link elements have an initial length equal to zero. Which node means I have added the displacement in the isolator?
Thanks again.
I have 74 isolators in my model, it is a super-long isolated structure with irregular plane. I want to do dynamic history analysis.
I still wonder that if I develop a simple sine wave, should I add this real time history on node1 or node2? Since all link elements have an initial length equal to zero. Which node means I have added the displacement in the isolator?
Thanks again.
Re: pre-displaced isolator
Fiona, if you examine the isolator hysteretic curves/data in the post-processor you will be able to determine the displacement in the isolator at the end of the sine wave.
Tim Huff
-
seismosoft
- Posts: 1263
- Joined: 06 Jul 2007, 04:55
Re: pre-displaced isolator
Hi Fiona and Tim,
Applying an initial displacement at the isolator nodes will fix the displacement to the initial value throughout the analysis.
Hence, the bottom isolator node (fixed to the ground) will move according to the ground displacement (derived from the integration of the acceleration record), whereas the top node (connected to the structure) will be fixed at a constant value. This will cause very large artificial displacement in the link element that models isolator, espeically when the record is not corrected.
Some ideas on how to achieve the initial deformation that you want
(1) Tim's idea is probably the best
(2) use a record that runs before the actual record that you want to analyse, and causes a permanent deformation. Maybe run an IDA analysis before (in a trial and error procedure), so that you measure at which loading factor you actually get the sought deformation
(3) The same effect might be achived by a horizontal force applied at the top node of the link element. This is the least desired method, because the horizontal force does not incorporate all the effects of a previously created residual deformation (e.g. moments from P-delta effects)
SeismoSoft Support Team
Applying an initial displacement at the isolator nodes will fix the displacement to the initial value throughout the analysis.
Hence, the bottom isolator node (fixed to the ground) will move according to the ground displacement (derived from the integration of the acceleration record), whereas the top node (connected to the structure) will be fixed at a constant value. This will cause very large artificial displacement in the link element that models isolator, espeically when the record is not corrected.
Some ideas on how to achieve the initial deformation that you want
(1) Tim's idea is probably the best
(2) use a record that runs before the actual record that you want to analyse, and causes a permanent deformation. Maybe run an IDA analysis before (in a trial and error procedure), so that you measure at which loading factor you actually get the sought deformation
(3) The same effect might be achived by a horizontal force applied at the top node of the link element. This is the least desired method, because the horizontal force does not incorporate all the effects of a previously created residual deformation (e.g. moments from P-delta effects)
SeismoSoft Support Team