Seismosoft | Earthquake Engineering Solutions

Hi,

I'm modelling a structure with a base isolation system. The isolation system is made by 20 friction pendulum isolators, which are represented by the link element plst. The problem is in the output results.I note that there isn't any acceleration at the base of the structure, at the level of the isolation system. How is it possible? Where is the error? I attach here my model, can you help me?

Thanks

https://www.dropbox.com/home/model

Hi a.sartori. The link you provide is asking for a password, so I wonder if you could e-mail the model?

It is unclear what you mean by "there isn't any acceleration at the base of the structure..".

What sort of analysis are you attempting? Dynamic time history analysis? IDA?

With a bit more information, perhaps the forum will be able to provide some insight a.sartori.

Regards,

huffte

Excuse me, huffte.

I'm attempting a Time History anlysis of a structure with a base isolation system. In the global responce parameters (on the post-processor), the acceleration of the node at the base of the structure during the Time History analiysis is 0.I think that there is an error in the model of the friction pendulum isolators. I model the friction pendulum in this way: the node of the base of the pillar of the structure above ground is connected to the node in the structure through a rigid link bound only in the z direction. Then, I created a node coincident with the node of the base of the pillar, and I bound in all traslational and rotational direction. Between this fixed node and the node at the base of the pillar I put the link element PLST. It may be a correct formulation?

Username and password to see the model are:
username or email address: a.sartori1988@gmail.com
password: Scapolan65
https://www.dropbox.com/home/model

Hi a.sartori.

I have looked over your model. This is certainly an ambitious endeavor you have taken on.

It seems highly unusual to apply different time history curves at various points along the length of all piles in the structure. It would seem more reasonable to apply the ground motion time histories at the base level of the structure. This is one recommendation, which you may or may not be willing to incorporate depending on your goals for this analysis.

Also, as the model currently exists, all acceleration loads are applied to the piles and the base - located between the piles and the ground floor - is fully restrained. So I am not surprised that the isolators are seeing no load. In fact, I suspect that the only thing being loaded is the piles with the current restraint strategy.

I would recommend not only revisiting the way in which the time history of acceleration is applied to the structure, but also the restraint strategy, a.sartori.

I have e-mailed a sample model of mine which consists of FPS isolators. I hope it might assist you in developing your model. It is quite complex.

Best of luck.

Hi,
Are friction pendulum isolator modeling results by plst link with experimental works have been verified?

Hi mohammadzadeh,

Upon reading your post, I quickly googled the expression "modelling friction pendulum systems" and found a large number of seemingly pertinent publications, amongst which I am sure there will be some that discuss the adequacy of using a relationship of the 'plst' type to model FPS.

In alternative, of course, you may of course carry out such numerical-experimental-analytica comparisons yourself.

Regards,

Rui

Dear Rui

Thank you for your suggestion.I read some experimental results and i believe that "plst link" can not adequate for FPS modeling verification, beacuse "plst link" application is based on a simplified modeling method(not accurate) and it does not consider friction coefficient variations with velocity and pressure.I attach a comparison of analytical and experimental shear force-bearing displacement loops.

https://www.dropbox.com/s/5eqo80bxgwdo6 ... n.jpg?dl=0


Best regards

mohammadzadeh,

Agree fully with you.

Still, I wonder, if one is modelling a multi-storey building isolated with a few tens of isolators, will the use of a precise FPS model lead to results that are much different from those obtained using an approximate FPS model?

I actually do not have the answer to the above question, since I never carried out such comparison myself, but I do wonder if the changes are significant - is there anybody out there who will have carried out such a comparative study?

Nonetheless, I repeat that I agree with you that it will be ideal if in the future SeismoStruct can feature such more precise FPS models (see my response to your post on the same topic in the 'Future Developments and Requests' section of this Forum).

Best,

Rui

hello everyone,

i am modeling a building with friction pendulum isolators, i performed an incremental dynamic analysis. i checked the results and recognised that there is a displacement at the base node which is fully fixed. the isolators are place on top of that columns. what am i doing wrong?

If you run dynamic time-history analysis the absolute displacement of a fixed node is the one derived from the integration of the acceleration time-history applied to that node. What is of interest to you is the relative displacement between the other nodes and the fixed node.

Seismosoft Support