STL issues

Submitted by andrea.pasquali on Tue, 05/15/2012 - 16:29

Hi all,
I know in LIGGGHTS 1 tria is like 11 particles, but I'm wondering if in LIGGGHTS there is a maximum number of Tria Elements or a Correlation between tria/particle size.
This because I run a case with particles size R=8-12 mm with a moving stl (num. elements = 36720, size = 2-5 mm) and when stl starts to touch particles it crashes with seg. fault!
(If I use a coarse / interm. stl mesh it works fine).

Another question regarding stl. Where is evaluated the momentum (Nm) on stl? Around the origin (0 0 0)? Is it possible to give to LIGGGHTS a different center of rotation?

Thanks

Andrea

Forums:

ckloss's picture

ckloss | Wed, 05/16/2012 - 10:47

Hi Andrea,

triangles should not be much smaller than the particle diameter. Currently the STL mesh implementation is serial, but there's going to be a revision of this soon.

>>Where is evaluated the momentum (Nm) on stl? Around the origin (0 0 0)?
There is a center of mass calculated which is the reference point

>> Is it possible to give to LIGGGHTS a different center of rotation
That's a good point, I'll add this to the next major version (2.0)

Cheers, Christoph

andrea.pasquali | Tue, 05/29/2012 - 09:09

Hi Christoph,
thank you for your reply!
I have another question regarding STL: is it possible to dump stresses on STL averaging them every N iterations direclty from LIGGGHTS? (as I do for forces)

Andrea

Silias | Wed, 05/30/2012 - 09:03

Hi Andrea,

yes, you can do that. You just have to output the stresses by f_ID[1 or 2 or 3 ...] and save these values in a variable.
Then you can do whatever you want with that variable - for example averaging them by fix_ave_time.

kind regards
Sebastian

andrea.pasquali | Wed, 05/30/2012 - 14:16

Hi Sebastian,
What you suggested is regarding the integral value in X Y Z for the ST?
Maybe I didn't explain well, I'd like to map stress on STL in VTK format not every dump-time but averaging it, for example, from iteration 0 to iteration 1000; or from the start to the end of simulation.
Do you know is it possible to do this?
Thanks

Andrea

Silias | Thu, 05/31/2012 - 09:40

Hi Andrea,
yes, that's what I meant.

Please explain, what results you want to postprocess (in paraview), I maybe don't understand what the problem is....

kind regards,
Sebastian

andrea.pasquali | Thu, 05/31/2012 - 12:16

Hi Sebastian,
I'm using this kind of commands:

fix cad_0 all mesh/gran/stressanalysis cad0.stl 2 1 0 0. -0.2 0. 0. 0.
fix forceOnCad_0 all ave/time ${Nevery} ${Nrepeat} ${Nfreq} f_cad_0[1] f_cad_0[2] f_cad_0[3] f_cad_0[4] f_cad_0[5] f_cad_0[6] file forces_0.dat
dump dumpCad_0 all mesh/gran/VTK ${dumptime} post/cad_0_*.vtk stress cad_0

With the first command I evaluate the stresses on stl. With the second command I'm averaging the integral of force writing it in a .dat file while with the third one I'm dumping the STL cad in VTK format every dumptime.
What I want is to average the stresses on STL cad and write them in VTK file.
Should I try something like:

dump dumpAveCad_0 all mesh/gran/VTK ${dumptime} post/cadAve_0_*.vtk stress forceOnCad_0

calling in the dump command the "forceOnCad_0" instead "cad_0"? Will it work?

Andrea

ckloss's picture

ckloss | Mon, 06/04/2012 - 18:33

Hi Andrea,

I think if you would want to average the value that you write to VTK you would have to do it by either post-processing or hard-coding it in fix_mesh_gran_stressanalysis. If you have questions on how to do this, feel free to use the developers forum!

Cheers, Christoph