soil structure interaction

[mmc]

Hi again!
I have done what you mentioned in your post. However, these errors appear. I cant select a different type of region...just for elements...it seems that the command is still the same...
Let me know.
Regards

[STKO Team]

Hi, I think you changed the python files I sent you, while STKO was running.
Please save your document and close it. Then open it again and it should be there.
Let me know

[mmc]

Hi again!

I did that before posting and didnt work. I did it again and finally it is solved.
The model seems to be working. However, the analysis stops in the 50th step... not reaching convergence...I have changed all of the parameters of the pushover analysis but no different results are obtained.

Could you please check again the model and try to guess what else can I change?
Thanks
Regards.

[STKO Team]

dear User,
the fem model is fine.
You can try to check the input parameter (material, section properties and analysis) and also to reduce time step size.
Have a nice day.

[mmc]

Hi again STKO!
I have downloaded the new version of the software and tried to run the model I attached in my previous post. However, tt does not converge. I do not know which might be the problem... Could it be something related to the new version of the software? Does the new version include the updated python files that you posted previously?
Regards

[STKO Team]

The only difference is that now each analysis call the wipeAnalysis at the end. It is fine but it seems it interacts with the folloiwing displacement.-control analysis.

However I changed some parameters and now it seems to work.
First your penalty values where too low (1e8)
Then I chnaged the norm type = 0 (abs-max) because you are using the contact element that uses lagrange multipliers to enforce the contact condition, and it puts contact forces into the displacement solution vector

p10_8.zip

(321.12 KiB) Downloaded 333 times

[mmc]

Hi STKO team,
Yes, it iworks now so thanks for that.
Since it takes a lot of time, I have changed into the parallel extension. Now, my computer can use the four cores and the analyses are way more faster. I have been able to perform nonlinear static analyses changing the parameters of the analysis's step. Next thing to do is performing dynamic anlyses and for that, I do not know if I could use the parallel option. Also, I do not know which are the characteristics of the analysis's step. Do you have any example of how it can be done? Maybe in some of the tutorials that you have been doing?
I changed different things. For example, I modelled the foundations and applied Eqdof. The model worked perfectly. Another thing is that I wanted the model to performe as nonlinear and for that, I used the PressureIndependMultiYield Material instead of Elastic. Thing is that when using that material, the soil is not moving but the structure works fine... I have used the parameters available in the Opensees wiki. I do not know if I have to add any additional step in the analyses... What do you think?

[Clarabella]

Dear mmc,
please watch these webinars (https://asdeasoft.net/?stkowebinars-for-opensees) on NLTH:

#018 Learning Opensees: Nonlinear Time History Analysis
#010 Learning Opensees: Nonlinear Analysis of an existing RC building including RC joint modeling
#08 Learning Opensees: SSI (Soil Structure Interaction)

Regards

[mmc]

Thanks for your help.
I have performed a time history analysis following the procedure established in the webminar. The soil is moving...However, I havent found why the soil doesnt move in the pushover analysis...
For the time history analysis, I have used the parallel analysis. However, in the post-processor mode, the results only appear for one of the partitions... I do not know what might be wrong. You can see attached my model.

[STKO Team]

Dear user,

The soil is moving...However, I havent found why the soil doesnt move in the pushover analysis...

It happened because in the last loadPattern you did not include the faceForce condition.

However, in the post-processor mode, the results only appear for one of the partitions... I do not know what might be wrong

Thanks for finding this bug. This happens when you have beams on only some of the partitions. We have already fixed this problem and it will be released with version 2.0.0 next week.
In the meantime I used a trick to force STKO to put at least 1 beam of the structure to every one of the 4 partitions. I added a face interaction (without any physical property or element property) that links the beams to the 4 sides of the soil. In this way the partitioner thinks that the beams are connected to the soil sides, (while they are not in OpenSees)!

One more note about the PressureIndependMultiYield:
It is a material model that uses the "updateMaterialStage" to move from elastic to plastic. Have a look at the documentation:
https://opensees.berkeley.edu/wiki/inde ... d_Material

Here is the fixed file (you should add the updateMaterialStage if you want)

p10_16.zip

(689.59 KiB) Downloaded 337 times