Learning Forum

[Ruben Waeytens]

Hi

I am trying to model a granular phase being transported by 2 corotating screws. The distance between these screws dictates the usage of an overset mesh or dynamic mesh. A sliding mesh is not an option. I have opted for an overset mesh but then I read this in the manual about the compatibilities:

However, when I am setting up the case, I do get the possibility to select a granular phase, albeit without the possibility to select some of the submodels like the frictional viscosity. This is shown in the picture below:

Is it possible to use a granular model for 2 interrotating screws with the overset mesh? If not, which other method do you propose?

Thank you in advance for your response.

[Rob]

What is the goal of the model? Just thinking that Rocky with/without Fluent coupling might be an option. 

[Ruben Waeytens]

The goal is to model the transport of a powder by a pair of screws. The use of DEM is not an option given the number of particles that would be in the system. I want to use the kinetic theory of granular flow. But how about using the KTGF in an overset mesh configuration?

[Rob]

I suspect the overset data transfer doesn't include some of the granular terms. It was intended for certain applications, notably things falling off aeroplanes, and Eulerian Granular contains a lot of additional terms that are unique to that model. 

Particle count could be an issue in Rocky, but options include coarse graining. Purely in Fluent, can you mimic the granular phase as a liquid with nonNewonian viscosity? 

[Ruben Waeytens]

Thank you for your reply. I suppose that this lack of data transfer is not a problem for dynamic meshing?

[Rob]

Correct, but you'll need to be more careful with the gap between the screws to ensure the mesh remains valid. 

[Ruben Waeytens]

Is it possible to use a poly-hexcore mesh for this or does it need to be all tetrahedrals? From what I understand from the manual, the remeshing methods only work for tetrahedrals. I also don't find a guideline on how many cells should be in between a moving wall and a stationary wall for dynamic meshes. Could you help me with this? Thank you in advance!

[Rob]

Definitely not poly-hexcore, I'm not sure if poly cells can be remeshed at present as I rarely look at remeshing: I have colleagues who do that sort of thing. Tets definitely work so are a safe starting point. 

Cells per gap is a case of "it depends". You need to ensure the cell quality is good. If you only have 1-2 cells in a gap, you may find flow leakage isn't well resolved, but you then need to consider if that's important, and whether you can afford the cell count to improve the resolution.