The cyclic-boundary can't work in a 3D case

MFiX How to
,
#1

Hello, teachers. I built a 3D dem case to test the cyclic boundary’s effect. But the particles in this case didn’t seem to implement a cycle, I don’t know why, can you check it for me, thanks!
3d-cyclic_2023-11-05T141121.737993.zip (45.3 KB)

#2

The cyclic boundary does not support DMP calculation mode?

#3

Looks like you have conflicting BCs. You already have a MI at y=ymin and a PO at y=ymax, so you can’t have a cyclic BC in the y-direction.

#4

Thank you very much for your reply! I removed the MI/PO boundaries in the y direction and set the gas MI/PO in the x direction, but the y-cyclic boundary still didn’t work. Attached is my new case.
3d-cyclic_2023-11-07T101905.753938.zip (71.0 KB)

#5

Instead of the MI/PO at the bottom/top, you set NSW BCs so there is still a conflict. Please remove the NSW BC at the bottom (y=ymin) and top (y=ymax) planes.

#6

Thanks a lot. Your guidence is so helpful to me. Now the case normally works in the serial mode, but the sovler will crash in DMP mode. Is it true that the cyclic boundary cannot be computed in DMP mode? If true, it’s really bad news for me that I have to use cyclic boundary conditions at work.
3d-cyclic_2023-11-07T230414.174036.zip (2.3 MB)

#7

Cyclic BCs should work in DMP. Somehow it looks like the gas density is not properly initialized in corner cells. Please set the gas pressure in the IC so the density gets set.

#8

Thank you very much! It is running normally after setting pressure in the IC.