Particles insertion and restart

Submitted by andrea.pasquali on Fri, 04/01/2011 - 09:42

Hi,
I'm trying to set a volume filled with particles to use and re-use after.
My questions are:
1) What is the faster insertion method to fill a volume until the level I want?
I'm trying with ins pour/dev but I have to wait for the filling (and some time I have less insertion than wanted); with lattice I have to wait the KinEn decrease (and it doesn't respect the level I want). pour/dev/packing could help me?
2) After I filled my domain I want to use a fix movecad command. I tried two different runs: first, I run ins + movecad and obtain a result; second, I run restart movecad from previous ins and I obtain a different result (monitoring the pressure on cad). Why this? Am I missing something in restart?

Thanks for any reply

Andrea

Forums:

JF's picture

JF | Fri, 04/01/2011 - 10:25

Hi Andrea,

> 1) What is the faster insertion method to fill a volume until the level I want?

You need to acquaint oneself with the simulation in.packing in the folder: examples/LIGGGHTS/packing.
It is an example of a radius expansion algorithm until the density (solid fraction) that you want.

Regards
JF

andrea.pasquali | Fri, 04/01/2011 - 15:44

Hi JF,
thank you for your reply.

I tried the in.packing. It fills the entire volume increasing the particles radius, right?
I tried it with my model (following the in.packing file):

fix pts all particletemplate/sphere 1 atom_type 1 density constant 1500 radius constant 0.0025
fix pdd all particledistribution/discrete 1. 1 pts 1.0
variable alphastart equal 0.20
variable alphatarget equal 0.60
variable growts equal 50000
variable growevery equal 40
variable relaxts equal 100000
region bc block -0.05 0.05 -0.05 0.05 -0.40 0.0 units box
fix ins nve_group pour/dev/packing 1 distributiontemplate pdd vol ${alphastart} 200 region bc

But my domain remains unfilled until the level I want to.
I have some questions:
1) is it possible to define a range of grow radius for the same atom type? For example as I said: in the end I want particles with radius between 2 and 5
mm.
2) is it possible in ParaView view the particles keeping the exact radius value? (maybe with Point Sprite plugin?)
3) I did not understand well the vol option, how it works during the iterations

Thanks

Andrea

JF's picture

JF | Sun, 04/03/2011 - 17:07

Hi Andrea,

> 1) is it possible to define a range of grow radius for the same atom type? For example as I said: in the end I want particles with radius between 2 and 5

This algorithm is largely used by the DEM users, I think that it is possible to define your grain size (load your own data file) at the density you want (if you accept overlaps). However, I am not anymore familiar with it into LIGGGHTS, you need to play more with this simulation to understand its features.

> 2) is it possible in ParaView view the particles keeping the exact radius value? (maybe with Point Sprite plugin?)

It is possible with Glyphs, you have to chose in the Glyphs properties:

Scalars: radius
Glyph type: Sphere
Radius: 1.0
Scale Mode: scalar
Set Scalar Factor: 1.0

You have to untick: Mask points and Random Mode

> 3) I did not understand well the vol option, how it works during the iterations

I think it is this one : vol ${alphastart}
Maybe, it is just to start with a sphere packing whith a density equal to 0.25.

Regards
JF

andrea.pasquali | Tue, 04/05/2011 - 10:02

Hi JF,
to view particles keeping the radius you can use also the Point Sprite plugin
Point Sprite Mode: Sphere
Scale by: radius
Edit radius transfer function: untick Use Scalar Range and tick Proportionnal

This is faster and better to view than glyph

regards

Andrea

JF's picture

JF | Tue, 04/05/2011 - 11:19

Hi Andrea,

You are right, the Point Sprite is better. However some video cards do not support the point sprite, it is my case.
So, I got used to visualizing my simulations with Glyph filter. It is not a good habit ;-), sorry !!!

Regards
JF

ckloss's picture

ckloss | Mon, 04/04/2011 - 09:11

>>2) I run restart movecad from previous ins and I obtain a different result (monitoring the pressure on cad).
>>Why this? Am I missing something in restart?

Is this a significant difference? with or without a fix ave/time?
Christoph

andrea.pasquali | Mon, 04/04/2011 - 09:28

Hi Christoph,
I checked the integral of pessure on .vtk by ParaView.
With insertion + movecad I have 16.866 N, with movecad restart I have 15.318 N.
I did not use the fix ave/time yet.
I will use it and check again.

Thanks

Andrea