CouplingProperties meanings for cfdemSolverIB solver

Submitted by paulaalejandrayo on Thu, 11/17/2016 - 17:29

Hello, This is the firts time I write here, I am starting to work with CFDEM and I am trying to figure out the best way (the most accurate and optimized) to set up my simulation. I am trying to set up a dense system of particles and an inlet for fluid injection from a side.
I am using the run/tutorials/cfdemSolverIB/twoSpheresGlowinskiMPI/CFD/contant as reference for this case (I want to have the fluid behaviour as well as the particles/particles interactions). I am using cfdemSolverIB solver.
I was trying to understand couplingProperties meaning for some enties of this case and I am not sure about
a) What is the difference between voidFractionModel "bigParticle" and voidFractionModel "IB" ? When will it be better to use one or the other?
b) What does minimumVoiFraction mean? is it in the minimum void fraction for the complete simulation domain (i.e. macroscopic) or it is about void fraction in each CFD cell?
c) What does zeroGrad mean for the voidFraction field initial cnodition (in CFD/0 folder) and it is possible to set up different boundary condition for this?
d) in run/tutorials/cfdemSolverIB/twoSpheresGlowinskiMPI/CFD/contant/dynamicMeshDict example) it says
"field interFace; " what field is this?
e) in the same file, does lowerRefineLevel/upperRefineLevel correspond to the absolute magnitude of the field or it is a proportion of this? I mean, should I scale those using the values of the field I am setting (for example U=10 in the inlet so lowerRefineLevel=1 and upperRefineLevel=9 for example?)
I will appreciate any help. Sorry if they are so many general question but I am just starting. Thank you in advance.

j-kerbl's picture

j-kerbl | Wed, 11/23/2016 - 11:59

Hi,

a) If you use the IB solver it is best to use the voidFractionModel IB, since it is able to completely cover cells with particle volume (void-fraction=0). The big-particle is used with other solvers, when the equations do not allow the void-fraction to be 0.
b) It is the minimum void-fraction for the cells. The reason for this is meantioned above.
c) zeroGradient means the face value at the boundary is the value of the center. Since void-fraction is calculated explicitely and not solved via a differential equation, the impact is rather small. You can e.g. set the void-fraction to be 1.
d) oviously the fluid-solid interface which needs to be resolved. The field is generated by the solver. Let the case run and look for it in the processor directories except at time=0
e) Yes the absolute values. https://openfoamwiki.net/index.php/Parameter_Definitions_-_dynamicRefineFvMesh

Cheers,
Josef

paulaalejandrayo | Thu, 11/24/2016 - 15:19

Thank you very much for your answer. In that case I should set the minimum void-fraction=0 to allow cells to be fully covered with particles, right? also, I am not sure I completely understand the meaning of the interFace field. for example, what does it mean in this case to set lowerRefineLevel=0.001 and upperRefineLevel=0.99 in the dynamicMeshDict? Thank you very much again.

Best,
Paula

cgoniva's picture

cgoniva | Fri, 12/02/2016 - 09:23

Hello Paula,

the dynamicMesh functionality is looking at the provided field (its name is specified in the dynamicMeshDict) and will do refinement for those cells where the field value is in between the border given (lower and upperRefineLevel).

Best regards,

Christoph Goniva

Arghavani | Sat, 03/14/2020 - 11:55

Hi Paula,
I am running the case cfdemSolverIB with some changes in the domain. actually, I have two semi-cylinder on two sides of the geometry and when the particles are reaching to this curve part some parts of particles are outside the domain and after some time the whole particles are backing again in the domain. But I want to have the whole volume of the particles inside the domain in the curve part as well. what I have changed in liggghts initial file is the boundary considered as m m m. but in tutorials is f f f. and when I am running the case with this case with a fixed boundary I get the error that some points are cut and it is reasonable since they do not exist. could you please give me a hint.
Best regards,
Arghavan

Arghavani | Sat, 03/14/2020 - 11:58

Hi Christoph,

I am running the case cfdemSolverIB with some changes in the domain. actually, I have two semi-cylinder on two sides of the geometry and when the particles are reaching to this curve part some parts of particles are outside the domain and after some time the whole particles are backing again in the domain. But I want to have the whole volume of the particles inside the domain in the curve part as well. what I have changed in liggghts initial file is the boundary considered as m m m. but in tutorials is f f f. and when I am running the case with this case with a fixed boundary I get the error that some points are cut and it is reasonable since they do not exist. could you please give me a hint.
Best regards,
Arghavan