Hi,
I am trying to simulate rocks that tend not to roll, is there a way that I can model translating only particles?
I tried using the rolling friction model with a high value of rolling friction;
pair_style gran/hertz/history rolling_friction cdt
with
fix m6 all property/global coefficientRollingFriction peratomtypepair 1 0.5
But the particles seem to rotate about the same value (comparing omega) no matter what I put in for the rolling friction value. Maybe I could just use an enormous value (greater than 1.0) for rolling friction? But I question the validity of the model at that point.
I have attached my input files (I think that the file extension needs to be changed back to .tar.gz).
Any help is greatly appreciated.
Thank you,
| Attachment | Size |
|---|---|
 Mill_11_12.tar_.gz | 270.49 KB |
kjkostlan | Thu, 08/16/2012 - 09:14
I needed to change the c++
I needed to change the c++ code a bit, my fix is in the developer forums under
"fix_move.cpp with a couple useful commands", remove the _txt decorations at the end fo the attached file's name. I can't attach files here.
It needs to be tested, but it probably works because it was a simple change.
RocheDEM | Thu, 08/16/2012 - 21:15
Can I use this fix_move in
Can I use this fix_move in version 2.0.6? Or do I have to install 1.5.3?
Do you have a small input script using your modified fix_move that I could "test drive".
Is there no other way to simulate non-rotating spheres???
Thanks
RocheDEM | Fri, 08/17/2012 - 05:04
What about just using fix nve
What about just using fix nve instead of fix nve/sphere...
I just tried it out, it appears to do exactly what I am looking for, the particles translate but do not rotate.
Any thoughts, is there a reason why I shouldn't use this fix?
kjkostlan | Fri, 08/17/2012 - 21:48
Did you check to make sure
Did you check to make sure that there is still sliding friction between the particles?
KFDEM | Sat, 08/18/2012 - 01:40
Yes, the particles for a pile
Yes, the particles for a pile with an obvious angle of repose. I use a differnt discrete element program at work (comercial), this program only allows translating particles (3DOF). Using fix nve in liggghts give me the same behaviour as the comercial code.
Give it a try, let me know what you think.
sbateman | Fri, 08/17/2012 - 17:59
You could try setting
You could try setting coefficientFriction = 0, which will cause F_tangential = 0 and so no torque will be applied to the particles.
Of course, that means you would be simulating completely smooth frictionless rocks, which is probably a bit unphysical.
ckloss | Mon, 08/20/2012 - 23:15
Ok, I'll have a quick look...
Ok, I'll have a quick look...
Cheers, Christoph
ckloss | Wed, 08/22/2012 - 12:08
Hi Kirk, the case is a bit
Hi Kirk,
the case is a bit too complex to handle it on a "quick look" basis. Can you track down the issue to a smaller scale where the effect that you see is easier to track?
Christoph
KFDEM | Wed, 08/22/2012 - 19:59
Hi Christoph, Next week I
Hi Christoph,
Next week I will put together a simple case, I will post it then.