Dynamic Group Assignments

Submitted by jtholle on Fri, 04/12/2013 - 00:36

Hello all,

I'm interested in simulating a horizontally rotating bin where particles leave through an orifice geometry that I have designed.

I'm mostly interested in how the particles interact with the orifice I have created as mesh geometry. However, I am also interested in the final distance the particles will travel and land due to gravity. Because I'm looking at a large number of particles, I would like to turn off particle interaction after the particles have reached a certain boundary after they leave the bin and allow them to move without interaction. They would fall due to gravity until they reach a specific horizontal plane (ground). The particles within the region containing the bin would still be undergoing both particle/particle and particle mesh interaction.

1) Is it possible to have a dynamic group assignment and generate a group that always contains all the particles that have left the region I'm initially interested in. If this is possible, I feel like I would be able to then exclude them from the interaction?

Alternatively,

2) Is it possible to write out the velocity, mass, etc of the particles as they leave the region so I could use some type of external program to propagate them in time?

3) Can anyone think of another technique that might allow me to do something like this.

I'm using Gran/hooke/history pair_style

Thanks,

Jeromy

Forums:

cstoltz | Fri, 04/12/2013 - 21:03

Do you need to track the particles after they hit the horizontal plane? If not, why not just have the position of that plane be the bottom boundary of the simulation domain and allow the particles to leave the simulation (thermo_modify lost ignore). It would speed up the simulation as you'd have fewer particles to track.

As for the analysis, I would just dump out particle positions for all particles and then, in the postprocessing script, identify those particles in regions of interest. I'm don't know of a method to create a dynamic group assignment (anyone else know how to do this?). Dumping and analyzing all of the particles will slow down the analysis somewhat, but the post-processing probably still won't take anywhere near as long as the simulation itself.

Regards,
Chris

ckloss's picture

ckloss | Fri, 04/19/2013 - 15:57

Hi Jeromy,

>>1) Is it possible to have a dynamic group assignment and generate a group
>>that always contains all the particles that have left the region I'm initially interested in
dynamic is not possible, but you can update the group assignments periodically during the run

>>2) Is it possible to write out the velocity, mass, etc of the particles as they leave the region
sure, you could just write out the data and then use delete_atoms to remove them

Cheers
Christoph

hunger's picture

hunger | Tue, 11/03/2015 - 13:40

Hi Christoph,

how is it possible to update the group assignments periodically in LIGGGHTS? I have thought about something like an if-condition in combination with the step-variable, but I did not manage to get the code to work.

Best regards
Harald

aaigner's picture

aaigner | Wed, 11/04/2015 - 10:23

I am not sure, which approach Chris was talking about, but one would be to use the every option of the run command.

Your command may look like
run every 1000 "group gName region reg" "do something here with the group" "delete_atoms group gName"