DEM mesh cells size

MFiX
#1

Hello everyone,

I’m currently working on setting up a DEM simulation in MFiX and have a couple of questions regarding DEM mesh cell sizing and their impact on computational efficiency.

  1. About the mesh size: I understand that the mesh cells need to be smaller than the particles but still large enough relative to the overall geometry of the system. But within this range, how does the size of the mesh cells influence the computation time? Is there a recommended approach to balancing mesh resolution with computational efficiency?
  2. When parallelizing: When using DMP, it’s mentioned that the number of DEM mesh cells should be at least three times the number of DMP mesh cells. Are there best practices or guidelines for determining proportional numbers of cells between DEM and DMP meshes to ensure optimal performance and accuracy?

I appreciate any insights or recommendations from your experience!

Thank you!

#2

Hello,
I am also a DEM learner, and share some experience with you.
firstly, I think the mesh size should be at least 3 times larger than the particle size to confirm the accuracy of DEM simulations. If the mesh size is smaller than the particle size, then the algrothm could be catagoried into direct numerical simulation (DNS).
sencondly, there is some guidelines about the grid partition for DMP runs in MFiX documents 13.9.2.

1 Like
#3

Thank you very much. I finally noticed that 3 times the particle size was the best ratio and MFiX documents 13.9.2. is indeed very usefull.