LIGGGHTS Simulation run problem

Submitted by Gulniamat on Sat, 02/27/2016 - 21:45

Hi every one....!
i am trying to simulate a mixer that contain glass beads with impeller in the middle.I am a new user of LIGGGHTS, i don't know how to make geometries for this setup.I means which components to be made in CAD?? which CAD software is to be used ???currently i made a vessel and impeller in PRO E and saved in STL format.when i use this then vessel geometry does'nt import nd i get problem with this.

I attached my input file also here,please any who have idea about this,kindly help me

atom_style granular
atom_modify map array
boundary f f f
newton off
communicate single vel yes

units si

region void cylinder z 0. 0. 0.13 0. 0.33 units box
create_box 2 void

neighbor 0.003 bin
neigh_modify delay 0

#Material properties required for new pair styles
fix m1 all property/global youngsModulus peratomtype 5.e6 5.e6
fix m2 all property/global poissonsRatio peratomtype 0.25 0.25
fix m3 all property/global coefficientRestitution peratomtypepair 2 0.5 0.3 0.3 0.
fix m4 all property/global coefficientFriction peratomtypepair 2 0.5 0. 0. 0.
fix m5 all property/global coefficientRollingFriction peratomtypepair 2 0.01 0. 0. 0.

#New pair style
pair_style gran model hertz tangential history #Hertzian without cohesion
pair_coeff * *

# Integrator

fix integrate all nve/sphere
# Gravity
fix grav all gravity 9.81 vector 0.0 0.0 -1.0

# Import mesh from cad:
fix cad1 all mesh/surface file meshes/vessel.stl type 2 curvature 1e-5
fix cad2 all mesh/surface file meshes/impeller.stl type 2 curvature 1e-5
fix mixer all wall/gran model hertz tangential history mesh n_meshes 2 meshes cad1 cad2

# physical characteristics of the particles:
fix pts1 all particletemplate/sphere 30492 atom_type 1 density constant 2456 radius constant 0.0015
fix pts2 all particletemplate/sphere 30492 atom_type 1 density constant 2456 radius constant 0.001
fix pdd1 all particledistribution/discrete 1. 2 pts1 0.5 pts2 0.5

# area where the particles are introduced from is set:
group nve_group region void
region factory1 cylinder z 0. 0. 0.11 0. 0.3 units box

# defines how the insertion is going to be done:

fix ins nve_group insert/pack seed 30000 distributiontemplate pdd1 &
insert_every 1000 overlapcheck yes all_in yes vel constant 0.0 0.0 -1.0 &
volumefraction_region 0.30 region factory1 ntry_mc 10000
# Timestep
timestep 0.00005

# Thermo settings
thermo_style custom step atoms ke cpu
thermo 20000
thermo_modify norm no #lost ignore
compute_modify thermo_temp dynamic yes

#insert the first particles so that dump is not empty
fix ctg all check/timestep/gran 1 0.01 0.01
run 1
unfix ctg
# Run 1 second to insert the particles
run 2000 upto

Forums:

cstoltz | Wed, 03/02/2016 - 16:17

You can make some basic primitive shapes from within LIGGGHTS, but generally it's easier to use a CAD package. Shouldn't really matter which CAD platform you use. You can save your geometry directly to STL format and try to import, but CAD packages typically save STL files using the edges of the geometry, which can lead to some really high aspect ratio triangles and cause problems in LIGGGHTS. You may have better luck saving the geometry to file, then using another application to create a surface mesh akin to a CFD simulation.

Regards,
Chris

Gulniamat | Wed, 03/09/2016 - 07:33

Thanku Chris ...
Kindly explain the last sentence "You may have better luck saving the geometry to file, then using another application to create a surface mesh akin to a CFD simulation. ",I failed to understand this.

msandli | Fri, 03/11/2016 - 22:40

for example, i create 3D geometries in Solidworks. Solidworks can save files in several different formats, including .stl. However, it is optimized to save some files better than others. Generally, Solidworks appears to try to use the least amount of triangles possible so it can use smaller sized files. While this can help on really large or complex geometries, this may not be ideal for the "simple" geometries that most people are probably using in LIGGGHTS. For example, a rather long cylinder saved as a .stl file by Solidworks will produce a body made up of many extremely tall triangles. where the base is at one end of the rod, and the opposite point is at the other end of the rod. If the cylinder is very long, these triangles will be extremely long and skinny, and many non-CAD specialized programs (LIGGGHTS, cfd software, etc.) will have trouble using these long skinny triangles.

To fix this, many CFD meshing programs create triangle meshes that seek to make each triangle as close to equilateral as possible. This is to make numerically solving the transport equations much more accurate. So, while the filesize itself will be bigger, the surface can be more accurately represented using more triangles.

A popular free software to use for this purpose is gmsh. I haven't tried it myself, but it appears to be quite powerful. If I were to use this workflow, it would look something like this:

Create geometry in Solidworks/CAD program -> save as some geometry file (probably .x_t, .stp, or .iges) -> open file in gmsh -> perform surface meshing operations -> save as .stl file -> use in LIGGGHTS

Gulniamat | Thu, 03/17/2016 - 19:08

kindly also tell me when i run the code and see the result in paraview its shows particles just like dots with spaces in between them.i want to see the particles in packed form means with no spaces as we see in real situation,is there any problem with code or its the problem with paraview setting or it always shows like this?

Gulniamat | Tue, 03/15/2016 - 21:46

Hi every one..!
i am trying to simulate a mixer that contain glass beads with impeller in the middle. I want to insert large particles first then small particles and then wanna to see mixing process. I used following input script but it has some problems like it shows only one type of particles in paraview. I don't know from where the problem is coming, is it the wrong use of run statement or some thing else ??? anyone plzzzzz kindly help me ,i worked on it for many days but could not succeeded. msandli nd Chris r u there??

atom_style granular
atom_modify map array
boundary f f f
newton off
communicate single vel yes

units si

region cyl cylinder z 0. 0. 0.09 0. 0.33 units box
create_box 2 cyl

neighbor 0.003 bin
neigh_modify delay 0

#Material properties required for new pair styles
fix m1 all property/global youngsModulus peratomtype 72.e6 180.e6
fix m2 all property/global poissonsRatio peratomtype 0.2 0.31
fix m3 all property/global coefficientRestitution peratomtypepair 2 0.94 0.5 0.5 0.97
fix m4 all property/global coefficientFriction peratomtypepair 2 0.4 0.5 0.5 0.42
fix m5 all property/global coefficientRollingFriction peratomtypepair 2 0.1 0.1 0.1 0.

#New pair style
pair_style gran model hertz tangential history #Hertzian without cohesion
pair_coeff * *

# physical characteristics of the particles:
fix pts1 all particletemplate/sphere 30492 atom_type 1 density constant 2456 radius constant 0.003
fix pdd1 all particledistribution/discrete 63243 1 pts1 1.

fix pts2 all particletemplate/sphere 30492 atom_type 2 density constant 2456 radius constant 0.002
fix pdd2 all particledistribution/discrete 45324 1 pts2 1.

# Import mesh from cad:
fix zwalls1 all wall/gran model hertz tangential history primitive type 1 zplane 0.0
fix zwalls2 all wall/gran model hertz tangential history primitive type 1 zplane 0.32
fix cylwalls all wall/gran model hertz tangential history primitive type 1 zcylinder 0.085 0. 0.
fix cad2 all mesh/surface file meshes/impeller.stl type 1 curvature 1e-5
fix mixer all wall/gran model hertz tangential history mesh n_meshes 1 meshes cad2

# Integrator
fix integrate all nve/sphere
# Gravity
fix grav all gravity 9.81 vector 0.0 0.0 -1.0
# Timestep
timestep 0.00005

# area where the particles are introduced from is set:
group nve_group region cyl
region factory1 cylinder z 0. 0. 0.085 0. 0.32 units box

# Thermo settings
thermo_style custom step atoms ke cpu
thermo 1000
thermo_modify lost ignore norm no
compute_modify thermo_temp dynamic yes
#insert the first particles so that dump is not empty
#fix ctg all check/timestep/gran 1 0.01 0.01
#run 1
#unfix ctg

# defines how the insertion is going to be done:

fix ins1 nve_group insert/pack seed 32141 distributiontemplate pdd1 verbose yes &
insert_every 1000 overlapcheck yes all_in yes vel constant 0.0 0.0 -1.0 &
volumefraction_region 0.17 region factory1 ntry_mc 10000

run 20000

# Particle insertion – large particles
fix ins2 nve_group insert/pack seed 62141 distributiontemplate pdd2 verbose yes &
insert_every 1000 overlapcheck yes all_in yes vel constant 0.0 0.0 -1.0 &
volumefraction_region 0.17 region factory1 ntry_mc 10000

run 20000

# Make a dump of particles and the stl file

dump dmp all custom 2000 post/dump*.mixer id type type x y z ix iy iz vx vy vz &
fx fy fz omegax omegay omegaz radius
dump dumpstl all mesh/stl 2000 post/dump*.stl

# Rotate central shaft
fix movecad2 all move/mesh mesh cad2 rotate origin 0. 0. 0. axis 0. 0. 1. period 0.3333

# Run 5 seconds
run 100000
unfix movecad2
run 30000

AnjanaKittu | Tue, 01/16/2018 - 15:54

I am having so much trouble getting my simulation running with 4 stl walls and around 30000 particles.
My issues: 1) Do not know how to get the stresses on walls
2) I have to run for a larger number of run, but I got the error "Invalid run N value" which I think needs multiple smaller runs. If anyone knows how to do it, please help me!
3) Finally, I need to get contact forces between the particles, and I use the following code. It doesn't seem to work. I am not sure if it needs a C++ coupling.
compute cf all pair/gran/local id pos force
dump cf all local 70000000 post/dump.fc.*.liggghts c_cf[1]

I am currently very much stuck with my work!!Any help is appreciated!!!

Thank you,
Anjana

richti83's picture

richti83 | Mon, 01/22/2018 - 20:15


My issues: 1) Do not know how to get the stresses on walls

https://www.cfdem.com/media/DEM/docu/fix_mesh_surface_stress.html

This fix stores a global vector with 9 components for access by various output commands. The first 3 components are equal to the total force on the mesh, the next 3 components store the total torque on the body exerted by the particles.

you can access these vector components with variable fx equal f_FIX-ID-OF-MESH-STRESS[1] and so on .. and than use it in fix print to output to file.

2) I have to run for a larger number of run, but I got the error "Invalid run N value" which I think needs multiple smaller runs. If anyone knows how to do it, please help me!

I think when you exceed maximum number of possible run steps your setup is wrong or the problem is not solveable with DEM (eg. to small timestep and to long needed simtime) maybe you can explain your setup in more detail, Are you aware that in DEM Young Modulus is reduced to reach larger timesteps ?

3) Finally, I need to get contact forces between the particles, and I use the following code. It doesn't seem to work. I am not sure if it needs a C++ coupling.
compute cf all pair/gran/local id pos force
dump cf all local 70000000 post/dump.fc.*.liggghts c_cf[1]

The command seems right (and needs no C++ coupling), but I'm wondering about this big dumpstep, does your simulation ever reach this iteration even with only two particles ? If not, no file is generated.
To force liggghts to dump at first iteration try following trick:
dump_modify cf every YOUR_DUMPSTEP first yes
the 'first yes' seems only to be allowed in conjuction with 'every', that's why one needs to redefine dumpstep too (some remains from LAMMPS I think).
But when there is no contact the file will only contain the header.

Best,
Christian.

I'm not an associate of DCS GmbH and not a core developer of LIGGGHTS®
ResearchGate | Contact