atom_style bond/gran

Submitted by zhou_yu on Thu, 05/11/2017 - 14:43

I noticed that atom_style bond/gran cant be recognized. The online documentation says "Note that bond/gran is an experimental code which is may not be available in your release of LIGGGHTS. An example for the sytnax is given below". I am wondering when it will be available to the public version user. Is this new atom_style for membrane boundary?

Ermek Asylbekov | Thu, 05/18/2017 - 14:47

Hi,
@mschramm: thank you for the link.
I've cloned and compiled it. My old cfdem simulation work out fine but I still have a problem with bond/gran.
I still get
log ../DEM/log.liggghts
thermo_log ../DEM/post/thermo.txt
atom_style hybrid granular ellipsoid bond/gran n_bondtypes 1 bonds_per_atom 6
ERROR: Invalid atom_style command (../atom_vec.cpp:72)

I compiled it as I did with the LIGGGHTS-PUBLIC version with
cfdemCompLIG
and
cfdemCompCFDEM

All variables are set correctly (tested with cfdemSysTest
Did I miss something?

thxia

Ermek Asylbekov | Fri, 05/19/2017 - 15:21

Thank you for your reply.
I've tried compiling liggghts on its own. I also went through the comlete installation process startig from setting up OF30x.
Still it doesn't work.
liggghts -help gives me:
LIGGGHTS (Version LIGGGHTS-PUBLIC 3.3.1, compiled 2016-01-14-14:02:34 by kloss, git commit 9b1fc57b77c483deaf2c0a6c69241c5d4f91e3c2 based on LAMMPS 23 Nov 2013)
Command line options:

-cuda on/off : turn CUDA mode on or off (-c)
-echo none/screen/log/both : echoing of input script (-e)
-in filename : read input from file, not stdin (-i)
-help : print this help message (-h)
-log none/filename : where to send log output (-l)
-nocite : disable writing log.cite file (-nc)
-partition size1 size2 ... : assign partition sizes (-p)
-plog basename : basename for partition logs (-pl)
-pscreen basename : basename for partition screens (-ps)
-reorder topology-specs : processor reordering (-r)
-screen none/filename : where to send screen output (-sc)
-suffix cuda/gpu/opt/omp : style suffix to apply (-sf)
-var varname value : set index style variable (-v)

Style options compiled with this executable

* Atom styles:

atomic charge ellipsoid hybrid line
sphere granular sph sph/var

* Integrate styles:

respa verlet

* Minimize styles:

cg

* Pair styles:

gran bubble gran_bubble hybrid hybrid/overlay
soft sph/artVisc/tensCorr

* Bond styles:

hybrid

* Angle styles:

hybrid

* Dihedral styles:

hybrid

* Improper styles:

hybrid

* KSpace styles:

* Fix styles

adapt addforce ave/atom ave/correlate ave/euler
ave/euler/stress aveforce ave/histo ave/spatial
ave/time box/relax buoyancy couple/cfd/convection
couple/cfd/speciesConvection couple/cfd/force
couple/cfd/dragforce couple/cfd/force/implicit couple/cfd
check/timestep/gran check/timestep/sph contacthistory
contacthistory/mesh deform/check deform diam/max
drag dt/reset efield enforce2d external
fiber/spring/simple freeze gravity
heat/gran/conduction heat/gran insert/pack
insert/rate/region insert/stream couple/lb/onetoone
lineforce massflow/mesh mesh/surface mesh/surface/planar
mesh/surface/stress mesh/surface/stress/servo momentum
move move/mesh move/mesh/gran multisphere
multisphere/nointegration neighlist/mesh nve nve/limit
nve/noforce nve/sphere nve/sph nve/sph/stationary
particledistribution/discrete particledistribution/discrete/numberbased
particledistribution/discrete/massbased planeforce print
property/atom property/atom/tracer property/atom/tracer/stream
property/global property/atomtype property/atomtypepair
region/variable rigid transportequation/scalar setforce
sph/density/continuity sph/density/corr
sph/density/summation sph/pressure spring spring/rg
spring/self store/force store/state particletemplate/multiplespheres
particletemplate/multisphere particletemplate/sphere viscous
wall/gran wall/region wall/region/sph wall/sph

* Compute styles:

atom/molecule bond/local centro/atom cluster/atom cna/atom
com com/molecule contact/atom coord/atom displace/atom
erotate/multisphere erotate/sphere/atom erotate/sphere
group/group gyration gyration/molecule inertia/molecule
ke/atom ke ke/multisphere msd msd/molecule
nparticles/tracer/region pair/gran/local wall/gran/local pe/atom
pe pressure property/atom property/local
property/molecule rdf reduce reduce/sph
reduce/region rigid multisphere slice stress/atom
surface temp

* Region styles:

block cone cylinder intersect mesh/tet
plane prism sphere union wedge

* Dump styles:

atom/vtk custom custom/vtk decomposition/vtk
euler/vtk image local mesh/stl stl
mesh/vtk mesh/gran/VTK movie xyz

* Command styles

change_box create_atoms create_box delete_atoms delete_bonds
displace_atoms minimize read_data read_dump read_restart
replicate rerun rotate run set
velocity write_data write_dump write_restart

So no sign of bond/gran

while compiling it works with bond_gran.cpp and bond_gran.d.
Why doesn't it recognize this style then?

mschramm | Sat, 05/20/2017 - 06:39

I don't know if it is something to do with the coupling. I do not have mine coupled with openFOAM. Once you have downloaded the liggghts with bonds from richti83, you should only have to go into src and compile.
Another thing that has me wondering is what you posted with the liggghts output. Please double check, but it shows that liggghts was compiled on January 14 2016...

Ermek Asylbekov | Mon, 05/22/2017 - 10:07

Hi,

you were right. LIGGGHTS compiled, but it didn't :D
So I deleted all files, cloned it again and compiled it with ubuntuVTK (I did it with fedora before, as I didi it with LIGGGHTS-PUBLIC version). Therefore I had to modify "Makefile.ubuntuVTK" in order to udate t to my VTK version. Afterwards it compiled properly and the bond/gran style is now recognized. Still I'm fighting CFDEMcoupling to recognize it aswell.
Wish me luck :D

Ermek Asylbekov | Mon, 05/22/2017 - 16:44

Hi,
I encounter the same error everytime I try cfdemCompCFDEMall or cfdemCompCFDEM:
In file included from /home/rahul/CFDEM/CFDEMcoupling-PUBLIC-3.0.x/src/lagrangian/cfdemParticle/lnInclude/dataExchangeModel.H:48:0,
from [..]/CFDEM/CFDEMcoupling-PUBLIC-3.0.x/src/lagrangian/cfdemParticle/lnInclude/momCoupleModel.H:44,
from [..]/CFDEM/CFDEMcoupling-PUBLIC-3.0.x/src/lagrangian/cfdemParticle/lnInclude/cfdemCloudI.H:33,
from [..]/CFDEM/CFDEMcoupling-PUBLIC-3.0.x/src/lagrangian/cfdemParticle/lnInclude/cfdemCloud.H:570,
from cfdemPostproc.C:44:
[..]/src/force.h:152:17: note: candidate: double LAMMPS_NS::Force::cg()
inline double cg()
[..]/src/force.h:152:17: note: candidate expects 0 arguments, 1 provided
Found this error a few times mentioned in this forum, but no solution.

mschramm | Tue, 05/23/2017 - 15:58

I don't know if any work has been done with this model to have it work nicely with openFOAM.
There very well may be extra code that needs to be written to make this work.
I am currently trying to fix a segmentation fault error and have the bonded model working in liggghts 3.6.x.
After this I could start looking at what is needed to get coupling working.

Ermek Asylbekov | Tue, 06/13/2017 - 14:28

I figured out that since the bonded model works with liggghts 3.3 I took an older cfdem version and it worked