Seismosoft | Earthquake Engineering Solutions

Dear Seismosoft,

I'm trying to model a friction pendulum seismic isolator with the link element plst. But i can characterize the curve with only two parameters: the force F0 and the stiffness K0. How can i define the maximum displicement of the isolators and the maximum force? In the postprocessor the hysteretic loop of the link element plst shows only the force F0, which is constant through the entire time history analysis and combined with very little displicement. Why? What may be due?

Hi a.sartori.

There is no need to define a maximum force/displacement for the PLST link element. You will need to examine these to determine the size of isolator to specify, which may or may not be within the range of those which have been fabricated and installed to date in real structures.

As far as the small displacement, it sounds as though the applied loading is such that the isolator never really "breaks free" and acts as intended given the properties specified, which is certainly possible. Perhaps you could toy with the specified yield force, in particular, to see if it "breaks away" at a smaller yield level.

Best of luck a.sartori.

I used the same "plastic curve" to model the pure frictoion connections beam-Pillar in an industrial prefabricated building.
The yield force F0 is exceeded many times during the history of acceleration but the displacements are very small, and I would expect them to be much larger.
I also tried to lower a lot of the yield force F0 but the movements are very small.
The same thing happens using a "symmetric bilinear curve" modelled like a curved plastic (very high initial stiffness, the same F0, the final stiffness anything).

How is this possible?

Well, I would say that it is quite possible. By definition, the plst link element yield at virtually any small displacement as long as the force in the element exceeds Fo.

What sort of analysis are you doing, lcasprini? Dynamic time history? If you bump up the load by specifying a factor double the actual value, do you get significantly larger displacements?

The fact that you get the same response when you try to mimic the plst element with a blsym element would seem to indicate that your model is responding as expected.

Best of luck lcasprini.

I tried to turn down to zero the yield strength F0.
I did a lot of tests and I never get high displacements.
Is not possible.

There is probably something I do not know about Non-linear dynamic analysis or about the seismic response of the curves (loading and reloading.

That does sound as if there may be some problem with the modeling strategy. If you would like to send the model, I will be happy to take a look at it as I find the time to do so.

I sent you an email.

Thank you.

Got it lcasprini. I'll start looking it over this weekend and early next week.

PLASTIC CURVE to MODELLING PURE FRICTION BEHAVIOR

After many tests carried out with very simple models (eg: simply supported beam) I am not able to reproduce the right behavior of "plastic curve".
With a time history nonlinear analysis the link does not "start" even if the yield strength F0 is exceeded.
This is not possible.
What am I doing wrong ?

I ask for help to the seismostruct support.

I have taken an initial look at your model lcasprini. You do have some unusual settings - for example, deriving loads from element masses with a z-factor of 9.81 which seems normal but an x-factor of 1 which seems odd to apply the element masses laterally. I do not believe that is the problem, but you should look into why exactly you have set the x-factor to 1. I agree with your assessment that there seems to be very little displacement in the links even after yielding. I will continue to examine the model. Perhaps others can join in too. Kind regards.