Question about another geoseismic modeling

Post Reply
youngkyu
Posts: 31
Joined: Tue May 04, 2021 3:36 pm

Question about another geoseismic modeling

Post by youngkyu » Wed Jun 09, 2021 7:43 am

Dear support team members,

I have another question about modeling seismic slope simulation.
As there is an issue with the simulation with free field columns linked to LK dashpots,
I am trying to model it with the fashion described in the below link (using free field column with 10,000m thickness)
https://opensees.berkeley.edu/wiki/inde ... s_of_Slope

However, when running my input file attached below,
the model did not work as desired.

When using the thickness of 5000m for free field columns, the vertical displacement contour shows that
slope part only displaced (free field column looks behaving as rigid body).

When using the thickness of 2m for ff columns,
the side of slope push the free field columns outward. (left and right)

Please take care of this issue as well.
Best,
Youngkyu
Attachments
RectSRA_PDMY_MFF.zip
(266.05 KiB) Downloaded 5 times

STKO Team
Posts: 660
Joined: Tue Oct 29, 2019 8:45 am

Re: Question about another geoseismic modeling

Post by STKO Team » Wed Jun 09, 2021 9:18 am

When you use the FF with a larger thickness (I set it to 10000), you also have to increase by the same factor:
  • Self Weight of FF
  • Base Action of FF
  • LK coefficients of FF
This is also described in the reference link of your previous post.

Another error was in your FixY condition (you forgot to switch off the Ux DOF)
RectSRA_PDMY_MFF.zip
(232.42 KiB) Downloaded 5 times

youngkyu
Posts: 31
Joined: Tue May 04, 2021 3:36 pm

Re: Question about another geoseismic modeling

Post by youngkyu » Fri Jun 11, 2021 8:50 am

Hi,

I would like to double check with you whether or not this model is correct.

This is because

when I ran the model with the tolerance of NormDispIncr equal to 1e-8, I got errors many times with adaptive time step, however, with the tolerance equal to 1e-6, I got the error a few of times.

So, I wonder if this model is correct or there is still something that I need to refine.

I have another question about creating *.out file by custom command.
When running the model with multi cores (4 cores),
the output file came out in an unorganized way.
Time vector appeared only at first and then a few of arrays were followed etc.

I wonder if there is something that I need to be careful for making the output file when running openseesMP ??

Best,
Youngkyu
Attachments
accNode.zip
(663.59 KiB) Downloaded 5 times
Slope_PDMY_MassFF.zip
(262.12 KiB) Downloaded 5 times

STKO Team
Posts: 660
Joined: Tue Oct 29, 2019 8:45 am

Re: Question about another geoseismic modeling

Post by STKO Team » Fri Jun 11, 2021 3:47 pm

when I ran the model with the tolerance of NormDispIncr equal to 1e-8, I got errors many times with adaptive time step, however, with the tolerance equal to 1e-6, I got the error a few times.
This is perfectly normal, the more strict is the tolerance, the more iterations OpenSees needs.
You just have to be careful to understand whether a tolerance of 1.0e-6 is fine for your model. If you are using meters then 1.0e-6 meters should be fine.

I have another question about creating *.out file by custom command.
When running the model with multi cores (4 cores),
the output file came out in an unorganized way.
Time vector appeared only at first and then a few of arrays were followed etc.

I wonder if there is something that I need to be careful for making the output file when running openseesMP ??
Standard OpenSees recorders do not work properly with OpenSeesMP.
This is because all OpenSeesMP instances (equal to the number of processors you are using) are attempting to write to the same file, and this is not allowed. So your files will end up containing garbage.
What you can do is to define the name of the recorder as a function of the current Process ID. In this case each process will write to a different file.

Code: Select all

recorder Node -file "MyOutput[getPID].out" ... 
However, keep in mind that in this way each file will contain the output related to only the nodes/elements that belong to that process.

Post Reply