restart file with no pouring

raguelmoon's picture
Submitted by raguelmoon on Fri, 10/15/2010 - 16:26

Hi LIGGGHTS users,
In new fix pour/dev, there is no pouring after i continue my simulation with restart file. Can anybody solve my problem?

Thanks in advance...
Ram

Forums:

ckloss's picture

ckloss | Fri, 10/15/2010 - 16:29

Have you re-specified the fix in the file you use for the restart?
Do you get a message like "restarting global state of fix XY", where XY is your fix pour/dev?
If no, please post the two scripts - otherwise its impossible to say

Christoph

raguelmoon's picture

raguelmoon | Fri, 10/15/2010 - 16:38

Hi Christoph,
Thanks for reply, I get the same message:

Particle insertion: 616.694444 every 115622 steps, 96000 within 1 steps
Resetting global state of Fix ins Style pour/dev from restart file info
Resetting per-atom state of Fix SHEAR_HISTORY Style SHEAR_HISTORY from restart file info
WARNING: One or more atoms are time integrated more than once
Setting up run ...
Memory usage per processor = 29.4002 Mbytes
Step Atoms KinEng 1 Volume
4000001 20000 50649.087 730.90042 386.76053
4001000 20000 49773.47 742.99934 386.76053
4002000 20000 48278.542 755.8848 386.76053
4003000 20000 46829.563 767.87408 386.76053
4004000 20000 45946.378 785.74867 386.76053
4005000 20000 44957.987 776.63818 386.76053
4006000 20000 43569.555 769.71968 386.76053
4007000 20000 42787.56 774.12031 386.76053
4008000 20000 41279.067 773.22958 386.76053
4009000 20000 40082.204 776.51568 386.76053
4010000 20000 38847.452 790.8771 386.76053
4011000 20000 38062.39 799.7289 386.76053
4012000 20000 37327.733 784.84428 386.76053
4013000 20000 36370.951 773.31578 386.76053
4014000 20000 35640.075 767.87976 386.76053
4015000 20000 34733.152 768.19486 386.76053
4016000 20000 33128.65 778.72526 386.76053
4017000 20000 31910.256 794.50325 386.76053
....
Thanks..
Ram

Ram

ckloss's picture

ckloss | Fri, 10/15/2010 - 16:44

Hi Ram,

post the scripts (both for init and restart) and I will have a look - which LIGGGHTS version have you used?

Christoph

raguelmoon's picture

raguelmoon | Fri, 10/15/2010 - 16:55

Hi christoph,
here is script :
....
...
fix pts1 all particletemplate/sphere 786 atom_type 1 density constant 2200 radius constant 0.06
#fix pts2 all particletemplate/sphere 1 atom_type 1 density constant 2200 radius constant 0.01
fix pdd1 all particledistribution/discrete 6778 1 pts1 1.0 #pts2 0.5

# Mesh from CAD

fix cad all mesh/gran/stressanalysis graphiteL2.stl 1 0.76 2.95 11.35 -3.16 90. 0. 0.
#fix cad2 all mesh/gran base_rec.stl 1 0.21 -1.195 0.5 -0.35 90. 0. 0.

# use the imported mesh as granular wall

fix granwalls all wall/gran/hooke/history 1 0 mesh/gran 1 cad #cad2

#region and insertion

#region bc block 0.005 3 0.005 3 16 18 units box
region bc cylinder z 2.3 2.3 1.49 16 17 units box
fix ins all pour/dev nparticles 96000 1 distributiontemplate pdd1 vol 0.08 10 vel uniform 0. 0. 0. 0. -3.0 region bc

#apply nve integration to all particles that are inserted as single particles
fix integr all nve/sphere

..
..
the same script in both cases but i simply increased the number of particles from 20000 to 96000 in fix pour/dev command.

Thanks..
Ram

Ram

ckloss's picture

ckloss | Tue, 10/19/2010 - 00:28

Hi Ram,

thanks for you report.

First - there was a minor bug in calculating the "end" time step when insertion is finished - that will be fixed in the next release.

The problem in your case is that the pouring was already finished in the run that you restart. This is in this case in contradiction to the # particles you specify in the input script.

As for your simulation, I would advise you to use a fresh fix pour/dev with a different ID from the one you used in the first script - there you can specify the # of particle you want to additionally insert.

I will add an error to the next release that hints the user to issue a fresh fix pour/dev if insertion was already finished in the restarted fix

Cheers,
Christoph

ckloss's picture

ckloss | Tue, 10/19/2010 - 09:09

Hi Ram,

after thinking about it, there may be a way of continuing pouring - I will change that in 1.1.9
PS: You have two integration commands in your scripts - is that intentional?

Christoph

raguelmoon's picture

raguelmoon | Wed, 10/20/2010 - 08:01

Hi,
Sorry, I put two integration commands in my script. I did all my simulation with that script. Can i continue simulation with restart file ( by removing one integration command in new script)? Will it effect the result or not?
Thanks...
Ram

Ram

ckloss's picture

ckloss | Wed, 10/20/2010 - 08:35

>>Can i continue simulation with restart file ( by removing one integration command in new script)
Sure

>>Will it effect the result or not?
It depends - the restart file you are reading was obtained with double integration.
If what you read in is an equilibrium situation, like static packing, the effect will not be too big.
If you are looking e.g. at free fall of particles - throw the file away

Christoph

raguelmoon's picture

raguelmoon | Wed, 10/20/2010 - 16:26

Hi,
In new version, if we want to pour more particles with restart file then we have to write n+m in fix pour/dev in new script. Here n is the number of particles in previous run and m is the number of particles we want to add. For example, we have poured 20000 particles and created restart file and in next run we want to add 5000 particles more, then we have to write 20000+5000= 25000, otherwise simulation will not run. Would it pour 5000 in new run or it would continue to pour 25000?
Thanks..
Ram

Ram

raguelmoon's picture

raguelmoon | Sat, 11/06/2010 - 02:20

Dear Christoph,
If I want to add more particles after the restart file in 3rd time then there is no pouring . If I re-specified the fix in the file then it pours again.
Best,
Ram

Ram