Problems with Simulation of continuous extruder

Submitted by aquinolarosa on Wed, 07/19/2017 - 03:10

Hi All,

I have a simulation of a continuous extruder. I noticed that as the simulation is progressing most of the particles sink to the bottom of the barrel/screw. Whilst the top view of the screw operates in a starved mode (less particles). I am not sure why the particles end up at the bottom. I have seen some simulation videos (see example attached) were this issue does not occur. Is there something I am doing wrong or if you have any suggests how I can improve my input script.

thanks

AttachmentSize
Image icon top view of screw97.12 KB
Image icon bottom view of screw105.34 KB
Image icon example from youtube113.43 KB
j-kerbl's picture

j-kerbl | Fri, 07/21/2017 - 13:50

Hi aquinolarosa,

not easy to say. But in comparison the example picture has a lot more particles per screw pitch section in comparison. Is this on purpose?
Maybe it needs a different initialization / insertion.

Cheers,
Josef

aquinolarosa | Mon, 07/24/2017 - 02:47

Hi Josef,

many thanks for your response. It is not possible to allocate how many particles should be in the screw pitch in the lights input script.
Even when I increase the number of particles I get the same issue. I have included the initialisation and insertion commands I use in my input script below. Please can you kindly check and advise me were I am going wrong in your opinion. Thanks in advance for your assistance.

###Initialise Simulation
#Particle for granular systems use sphere/granular
atom_style granular

#Data Structure to store
atom_modify map array

#f=non-periodic/fixed, in x,y,z
boundary f f f
newton off

#use single for granular systems
communicate single vel yes

units si

#Particle density and radius to be inserted
fix pts1 all particletemplate/sphere 15485863 atom_type 1 density constant 1540 radius constant 0.00045 volume_limit 1e-15
fix pts2 all particletemplate/sphere 15485867 atom_type 2 density constant 1114 radius constant 0.00013 volume_limit 1e-15

#Composition of the particle distribution
fix pdd1 all particledistribution/discrete 32452843 2 pts1 0.8 pts2 0.2

#Region and insertion
region factory block 0.001 0.008 0.001 0.029 -0.002 0.0 units box

#Particle insertion
fix ins all insert/rate/region seed 32452867 distributiontemplate pdd1 nparticles 90000 massrate 0.00027 insert_every 10000 &
overlapcheck yes vel constant 0.0 0.0 0.0 omega constant 0.0 0.0 0.0 region factory ntry_mc 100000

j-kerbl's picture

j-kerbl | Fri, 07/28/2017 - 14:26

Hi,

some things come to mind:
1.) In the example it looks as if the domain might be periodic.
2.) Are you applying any pressure at the outlet on the particles? Think about boundary conditions to mimic your flow regime.
3.) Do you get any warnings from the insertion during the run?
 

Cheers,
Josef

aquinolarosa | Mon, 07/31/2017 - 12:29

1) Yes the example is periodic. However, I am not trying to simulate one element as per the example. I would like to simulate the mixign of particles in along the whole screw.
2) No I am not applying pressure at the outlet on the particles.
3) I get the following warning: WARNING: Fix mesh: Mesh contains highly skewed element, moving mesh (if used) will not parallelize well (../surface_mesh_I.h:656)
INFO: Particle insertion ins: 12.361277 particles every 10000 steps - particle rate 494.451071, (mass rate 2.700000e-04)
90000 particles (mass 4.914541e-02) within 72800000 steps
Maybe I need to redo my mesh again?

thanks