Convergence issues in gravity with finer mesh

[konstantinosch]

Hello all,

I am trying to simulate a gravity initialization for a soil domain using SSP brick elements. Initially, I used a coarser mesh and the analysis was successful and the results as expected.

Then, I tried to increase the discretization of the mesh but the analysis does not converge.
I have tried different algorithms, tests and tolerances without succeeding.

Any help would be highly appreciated.

Find below the input files.

STKOresults.7z

(4.5 MiB) Downloaded 119 times

Thank you very much.
Kind regards,
Konstantinos

[STKO Team]

You should apply the constraintPattern before the analysis, otherwise during the analysis your model will be singular.
Also consider using parallel processing (openseesmp) with such a fine mesh:

testSTKO.zip

(2.89 MiB) Downloaded 138 times

[konstantinosch]

Hi,

Thank you for your prompt reply. Unfortunately, the model I sent you was re-created, and I forgot to change the sequence. However, after implementing your suggestion, the problem persists when using the OpenSees solver. With OpenSees SP, it was possible to proceed and obtain accurate results.

Additionally, I would like to simulate train-induced vibrations by applying nodal loads of triangular pulses across a line to simulate train movement. I have implemented this in OpenSeesPy with Python, but I was wondering if there is similar functionality in STKO.

Thank you.

Kind regards,
Konstantinos

[STKO Team]

So it's a problem of memory. If it works with OpenSeesSP and OpenSeesMP, but not with OpenSees, it's because your model does not fit in only one process.

Additionally, I would like to simulate train-induced vibrations by applying nodal loads of triangular pulses across a line to simulate train movement. I have implemented this in OpenSeesPy with Python, but I was wondering if there is similar functionality in STKO.

Depending on how you did it in Python, you can do the same using customCommand AnalysisStep in STKO, using an equivalent Tcl script.