Volume_Limit of fix/particletemplate

Submitted by NTT1508 on Tue, 01/20/2015 - 12:35

Hi all,

The manual describes about volume_limit of fix/particletemplate/sphere that LIGGGHTS will throw an error if the particle volume too small compared to machine precision. My question is how to adjust machine precision with default limit 1e-12 of LIGGGHTS. For examples, I'm playing with particle's size D=50micron that has volume of 5.4e-14 m3. It's obviously lower than default value. Does it cause serious error to simulation result?

Additionally, when I use fix/particletemplate to define particle's size then generate particles by fix/pack or fix/insert/rate, it warns about volume_limit issues however this warning does not appear when I create atom by create_atoms or data_read. Could anyone explain about this?

Thank for your concern,

Best regards,

Nathan,

Daniel Queteschiner | Wed, 01/21/2015 - 13:19

If your numbers are getting that small you may use CGS units instead of SI.

NTT1508 | Wed, 01/21/2015 - 23:46

Hi Daniel,

Thank for your hint much. Yes CGS can reduce scale of calculation and it would not exceed limit of machine. How If I do CFDEM coupling in that CFD requires SI Unit? The information exchange between CFD and DEM might be in conflict, I am afraid. Do I have any option to change Unit of CFD to be compatible with DEM using CGS. I think it is possible to add some unit scale when program exchange.

I have tried some running with CFDEM (SI unit) for my very small dimension (50 -1000micromet). But it is hard to check whether the result's good or not. Movement of particles looks quite normal.

I am looking forward to your advice,

Thanks,

Best regards,

Nathan,

NTT1508 | Mon, 02/02/2015 - 09:32

Hi Daniel,

Thank for your advice. I have tried with tutorial case ErguntestMPI_cgs. The graph comparing Analytical and numerical solutions (plotted by Octave) are matched. However when I try to run without copying controlDict file to .OpenFOAM, the result is the same ( no impact of Unit conversion file on result). So I have several confusions that would like to listen to your idea below.

1) Why running the same case with and without controlDict file not cause any difference?

2) Copying controlDict file from CFDEM/.../etc/ to OpenFOAM/.OpenFOAM is to add DimensionedConstants for cgs unit system. I am using OpenFOAM2.3.0 BUT in CFDEM/../etc there are only controlDict files for OpenFOAM 2.1.x and 2.2.x. Is that possible to use controlDict file of version 2.2.x for my current version 2.3.0 ? Hope there is no difference.

3) A part from Copying controlDict file, changing values of fields (U,p,nu,g...), Do I need to do anything else to completely convert Unit in CFD? I see another controlDict file in OpenFOAM/.../etc. Do we need to change this file too?

4) I create the same case with ErgunTestMPI_cgs but change unit to SI (change U,p,nu,g....). I check carefully to ensure all fields matched properly between cgs and SI. However, the result in both cases are not completely the same. A slight difference appears such as value of p_cgs/rho is 1477.86 while p_SI/rho is (0.147192) at the same point and time. 0.147192*10000=1471.92 compared to 1477.86 of cgs Unit. The same magnitude of differences occurs at other points. I think dimensions of particles in tutorial case is not really small (0.011m) so the result calculated in both Unit systems should be exactly the same. Do you think that is true or I made a mistake somewhere?

I am looking forward to your sharing.

Thank very much,

Best regards,

Nathan,

j-kerbl's picture

j-kerbl | Tue, 03/03/2015 - 17:12

Hi Nathan,

from what I see the controlDict file of the ErguntestMPI_cgs sets the units everytime you run it. It should not matter what the controlDict in the etc says. These should affect all the cases which have no specific setting of their own.
You can compare the settings of the controlDict in ErgunTestMPI_cgs and in OpenFOAM-2.3.x/etc/ and see if there any slight differences.

I hope that helps a little bit!

Cheers
Josef

NTT1508 | Wed, 03/04/2015 - 02:37

Hi Josef,

Thank for your response. Obviously I am aware of the differences between original controlDict and controlDict_cgs files. What I mean here is that copying controlDict_cgs to .OpenFOAM doesnot result in any difference from the case without this action.

Anyway, I noted that in new version CFDEM2.9.0, tutorial ErgunTestMPI_cgs, it has removed a command to copy controldict_cgs file to .OpenFoam directory. How come did it change?

Thank very much,

Regards,