Tutorial 3.3 does not run without specifying interpolation scheme

MFiX
1

If you follow the instructions as written in step 3.3.6 of tutorial 3.3. Two-dimensional fluidized bed, Discrete Element Model (DEM) — MFiX 25.2 documentation the code will not run. By default the field Interpolation is set to Field-to-particle and particle-to-field and the field Scheme is set to None. I suppose this combination does not work after reading this post: Interpolation for dem

Either the Interpolation field should be set to No interpolation or a scheme should be selected. I tested both and both work, although giving significantly different results. As a beginner I have no idea about what the Width of the DPVM scheme should be so I suppose it would be better to elaborate a bit on this option in the tutorial.

Furthermore, now that I’m digging into this I started wondering what I am actually doing when I: select Fully coupled in the field Coupling method, then set Interpolation to No interpolation, and finally set Scheme to None. The simulation does run and the results show moving particles. The comment that shows up when I hover over the Interpolation field says that interpolation is needed to calculate gas-particle interactions such as the drag force. So, if I understand correctly: shouldn’t Interpolation be set anyway, even if you have one-way coupling?

2

Thanks for pointing that out, we will update the documentation. We should set the default to “No interpolation” in the GUI and this will set the interpolation scheme to “None”.

When you select Fully coupled and set Interpolation to No interpolation, the drag is computed, but the gas velocity is taken at the fluid cell face center, not at the particle position. It is not required to use interpolation to compute drag.

1 Like
3

Thanks for your answer. Just for my understanding: which cell face centered velocity is used for computing the drag (in 3D we have 6 faces to choose from)? If a particle is in a cell somewhere and we don’t use any interpolation, but instead use the velocity at the fluid cell face center, I suppose the velocity at the closest face is used?

4

Sorry I wasn’t clear, we don’t directly use face-centered velocities to compute drag. We take the gas velocity at the face centers (there is a u-velocity at the east face and another one at the west face), and average them to get a cell-centered value. We repeat this for the y and z-components. All particles in a fluid cell use this cell-centered velocity to compute drag.

1 Like
5

That makes sense. Thank you so much for explaining @jeff.dietiker!