Particle insertion

Submitted by tkulju on Thu, 08/25/2011 - 15:23

Hi!
Is it possible to insert group of particles, so that their initial velocity would be against gravity field? I've tried it with

fix pour/dev

command, but it gives me an error message:
ERROR: fix pour: z-velocity must not not be <0.

Thanks,

Timo

Forums:

tkulju | Fri, 08/26/2011 - 10:00

Ok, I figured that out. By using the group and velocity commands I could set the initial velocity.

- Timo

ckloss's picture

ckloss | Fri, 08/26/2011 - 14:46

What you can also do (actually I would recommend that) is to use the newer fix insert/stream command where all assumptions about gravity have been dropped - it has a separate tutorial case

Christoph

tkulju | Mon, 08/29/2011 - 14:14

I'm having a little troubles with the fix insert/stream. For one stream it works fine, but for several streams it collapses (Segmentation Fault error).
Below is the snip of my code:

group ins_group region reg

fix ins_mesh1 all mesh/gran ins_face.stl 1 1.0 -0.080 0.291 -0.15 0.0 0.0 0.0
fix ins_mesh2 all mesh/gran ins_face.stl 1 1.0 -0.080 0.246 -0.15 0.0 0.0 0.0
fix ins_mesh3 all mesh/gran ins_face.stl 1 1.0 -0.080 0.201 -0.15 0.0 0.0 0.0

# first inlet
fix ins1 ins_group insert/stream seed 100001 distributiontemplate pdd2 nparticles 5000 &
vel constant 0. 0.0 1.0 particlerate 1000 &
overlapcheck yes &
insertion_face ins_mesh1 extrude_length 0.010

# second inlet
fix ins2 ins_group insert/stream seed 100002 distributiontemplate pdd2 nparticles 5000 &
vel constant 0. 0.0 1.0 particlerate 1000 &
overlapcheck yes &
insertion_face ins_mesh2 extrude_length 0.010

#third inlet
fix ins3 ins_group insert/stream seed 100003 distributiontemplate pdd2 nparticles 5000 &
vel constant 0. 0.0 1.0 particlerate 1000 &
overlapcheck yes &
insertion_face ins_mesh3 extrude_length 0.010

It usually crashes right after second inlet. I tried to unfix ins_mesh{1,2,3}, but it does not seem to help. Below is the output, if it's any help:


# second inlet
fix ins2 ins_group insert/stream seed 100001 distributiontemplate pdd2 nparticles 5000 vel constant 0. 0.0 1.0 particlerate 1000 overlapcheck yes insertion_face ins_mesh2 extrude_length 0.010
INFO: fix insert/stream: 'extrude_length' was provided, but is re-calculated using 'insert_every'
Particle insertion: 10.000000 particles every 20000 steps - particle rate 1000.000000 (mass rate 0.000524)
5000 particles (mass 0.002618) within 10000000 steps
[host:30372] *** Process received signal ***
[host:30372] Signal: Segmentation fault (11)
[host:30372] Signal code: Address not mapped (1)
[host:30372] Failing at address: 0xa8
[host:30372] [ 0] /lib64/libpthread.so.0 [0x30d720eb10]
[host:30372] [ 1] /opt/share/apps/LIGGGHTS/liggghtsdev/src/lmp_fedora(_ZN9LAMMPS_NS18FixPropertyPerAtomC1EPNS_6LAMMPSEiPPc+0x422) [0x566ad2]
[host:30372] [ 2] /opt/share/apps/LIGGGHTS/liggghtsdev/src/lmp_fedora(_ZN9LAMMPS_NS6Modify7add_fixEiPPc+0x1c58) [0x5cdf1a]
[host:30372] [ 3] /opt/share/apps/LIGGGHTS/liggghtsdev/src/lmp_fedora(_ZN9LAMMPS_NS6Modify24add_fix_property_peratomEiPPc+0x40) [0x5d0276]
[host:30372] [ 4] /opt/share/apps/LIGGGHTS/liggghtsdev/src/lmp_fedora(_ZN9LAMMPS_NS15FixInsertStream11post_createEv+0xa3) [0x52ff31]
[host:30372] [ 5] /opt/share/apps/LIGGGHTS/liggghtsdev/src/lmp_fedora(_ZN9LAMMPS_NS6Modify7add_fixEiPPc+0x3002) [0x5cf2c4]
[host:30372] [ 6] /opt/share/apps/LIGGGHTS/liggghtsdev/src/lmp_fedora(_ZN9LAMMPS_NS5Input3fixEv+0x1d) [0x5b3d87]
[host:30372] [ 7] /opt/share/apps/LIGGGHTS/liggghtsdev/src/lmp_fedora(_ZN9LAMMPS_NS5Input15execute_commandEv+0x52a) [0x5b5976]
[host:30372] [ 8] /opt/share/apps/LIGGGHTS/liggghtsdev/src/lmp_fedora(_ZN9LAMMPS_NS5Input4fileEv+0x2b6) [0x5b6f46]
[host:30372] [ 9] /opt/share/apps/LIGGGHTS/liggghtsdev/src/lmp_fedora(main+0x5f) [0x5be90f]
[host:30372] [10] /lib64/libc.so.6(__libc_start_main+0xf4) [0x30d661d994]
[host:30372] [11] /opt/share/apps/LIGGGHTS/liggghtsdev/src/lmp_fedora(__gxx_personality_v0+0x339) [0x46f419]
[host:30372] *** End of error message ***

- Timo

ckloss's picture

ckloss | Wed, 08/31/2011 - 16:27

Ok I think I know what is causing this. Check out the next version (1.4.2) to be released in the course of the next days and tell me if it helped.
Christoph

tkulju | Mon, 09/12/2011 - 12:05

Hi!
I installed the 1.4.3-version, but the fix insert/stream doesn't seem to work with multiple inlets. Now it gives an error:
ERROR: Error in fix property/atom. There is already a fix that registers a variable of the same name

Heres a snap of script:

fix face1 all mesh/gran face.stl 1 1.0 -0.080 0.201 -0.1 0. 0. 0.
fix face2 all mesh/gran face.stl 1 1.0 -0.080 0.249 -0.1 0. 0. 0.
fix face3 all mesh/gran face.stl 1 1.0 -0.080 0.291 -0.1 0. 0. 0.

group ins_group region reg

fix ins1 ins_group insert/stream seed 100001 distributiontemplate pdd2 nparticles 5000 &
vel constant 0. 0. 0.5 particlerate 1000 &
overlapcheck yes insertion_face face1 extrude_length 0.6
#insert_every 1000 &

fix ins2 ins_group insert/stream seed 100002 distributiontemplate pdd2 nparticles 5000 &
vel constant 0. 0. 0.5 particlerate 1000 &
overlapcheck yes insertion_face face2 extrude_length 0.6
#insert_every 1000 &

fix ins3 ins_group insert/stream seed 100003 distributiontemplate pdd2 nparticles 5000 &
vel constant 0. 0. 0.5 particlerate 1000 &
overlapcheck yes insertion_face face3 extrude_length 0.6
#insert_every 1000 &

Again, with a single inlet it seems to work.

- Timo

ckloss's picture

ckloss | Mon, 09/12/2011 - 14:40

Try replacing the files in the /src directory with the files I attached to this thread and recompile:
node/497

If it helps, I will include this in the next release

Christoph

tkulju | Tue, 09/13/2011 - 13:47

...and the error message for immediate crash is, if it's any help

WARNING: Dangerous build in triangle neighbor list.
[host:04252] *** Process received signal ***
[host:04252] Signal: Segmentation fault (11)
[host:04252] Signal code: Address not mapped (1)
[host:04252] Failing at address: 0x8
[host:04252] [ 0] /lib64/libpthread.so.0 [0x30d720eb10]
[host:04252] [ 1] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(_ZN9LAMMPS_NS23FixWallGranHertzHistory24deriveContactModelParamsEiddRdS1_S1_S1_S1_S1_+0x92) [0x5a1c1a]
[host:04252] [ 2] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(_ZN9LAMMPS_NS23FixWallGranHookeHistory13compute_forceEiddddddPdS1_d+0xfb) [0x5a29e3]
[host:04252] [ 3] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(_ZN9LAMMPS_NS11FixWallGran15post_force_evalILi0EEEvi+0x156b) [0x5a147b]
[host:04252] [ 4] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(_ZN9LAMMPS_NS11FixWallGran10post_forceEi+0x13) [0x59e307]
[host:04252] [ 5] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(_ZN9LAMMPS_NS6Modify10post_forceEi+0x46) [0x5c9ac2]
[host:04252] [ 6] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(_ZN9LAMMPS_NS6Verlet3runEi+0x3a8) [0x68ab74]
[host:04252] [ 7] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(_ZN9LAMMPS_NS3Run7commandEiPPc+0x668) [0x66ab6c]
[host:04252] [ 8] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(_ZN9LAMMPS_NS5Input15execute_commandEv+0xfb1) [0x5b717d]
[host:04252] [ 9] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(_ZN9LAMMPS_NS5Input4fileEv+0x2b6) [0x5b7cc6]
[host:04252] [10] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(main+0x5f) [0x5bf6ef]
[host:04252] [11] /lib64/libc.so.6(__libc_start_main+0xf4) [0x30d661d994]
[host:04252] [12] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(__gxx_personality_v0+0x339) [0x46f6c9]
[host:04252] *** End of error message ***

The difference is the different seed values (at least with seed-values 1,2001,10001).

- Timo

ckloss's picture

ckloss | Wed, 09/14/2011 - 16:49

Hi Timo,

thanks for the info. Can you please post the full script so I can reproduce the problem? I would then post a patch for the next version

Thanks!
Christoph

tkulju | Fri, 09/30/2011 - 10:05

Hi Christoph!
And sorry for the late reply. I suspect that there's some conflict between
insert/stream
and
check/timestep/gran
fixes. I commented out the
fix ts all check/timestep/gran 1000 0.1 0.1,
and there seems to be no problem anymore. I assume the files you attached are already in 1.4.4. version? Because that's what I'm using, and it's running fine.

- Timo

ckloss's picture

ckloss | Fri, 09/30/2011 - 10:44

Hi timo,

>> the files you attached are already in 1.4.4. version
yes

>>some conflict between insert/stream and check/timestep/gran
I don't think so... If you don't mind posting the script, I would have a closer look

Christoph

tkulju | Tue, 09/13/2011 - 13:37

Hi!
I don't know, is it relevant to this, but after some (990) iteration steps, I get the following error message

[host:04186] *** Process received signal ***
[host:04186] Signal: Segmentation fault (11)
[host:04186] Signal code: Address not mapped (1)
[host:04186] Failing at address: 0x10
[host:04186] [ 0] /lib64/libpthread.so.0 [0x30d720eb10]
[host:04186] [ 1] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(_ZN9LAMMPS_NS20FixCheckTimestepGran26calc_rayleigh_hertz_estimsEv+0x46f) [0x51e18f]
[host:04186] [ 2] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(_ZN9LAMMPS_NS20FixCheckTimestepGran11end_of_stepEv+0xe) [0x51e31a]
[host:04186] [ 3] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(_ZN9LAMMPS_NS6Modify11end_of_stepEv+0x5a) [0x5c9b88]
[host:04186] [ 4] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(_ZN9LAMMPS_NS6Verlet3runEi+0x3c5) [0x68ab91]
[host:04186] [ 5] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(_ZN9LAMMPS_NS3Run7commandEiPPc+0x668) [0x66ab6c]
[host:04186] [ 6] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(_ZN9LAMMPS_NS5Input15execute_commandEv+0xfb1) [0x5b717d]
[host:04186] [ 7] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(_ZN9LAMMPS_NS5Input4fileEv+0x2b6) [0x5b7cc6]
[host:04186] [ 8] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(main+0x5f) [0x5bf6ef]
[host:04186] [ 9] /lib64/libc.so.6(__libc_start_main+0xf4) [0x30d661d994]
[host:04186] [10] /opt/share/apps/LIGGGHTS/liggghtsdev-testing/src/lmp_fedora(__gxx_personality_v0+0x339) [0x46f6c9]
[host:04186] *** End of error message ***

Below is my insert fixes

fix ins1 nve_group insert/stream seed 1001 distributiontemplate pdd2 &
nparticles 5000 vel constant 0. 0. 0.5 massrate 2.158333e-3 &
overlapcheck yes maxattempt 2000 insert_every 10000 insertion_face face1

fix ins2 nve_group insert/stream seed 1002 distributiontemplate pdd2 &
nparticles 5000 vel constant 0. 0. 0.5 massrate 2.158333e-3 &
overlapcheck yes maxattempt 2000 insert_every 10000 insertion_face face2

fix ins3 nve_group insert/stream seed 1003 distributiontemplate pdd2 &
nparticles 5000 vel constant 0. 0. 0.5 massrate 2.158333e-3 &
overlapcheck yes maxattempt 2000 insert_every 10000 insertion_face face3

I looked at the post files until that point, and the particles were moving just fine. With pour/dev I haven't got this error message. And if I change the seed-value, it crashes right after the particles are inserted. Any idea, what's going on?

- Timo