Mesh with multiple boxes

MFiX How to
#1

I tried to create a geometry consisting of three rectangular boxes. When starting the solver I get the following error:
ERROR cartesian_grid/get_stl_data.f:674
ERROR: NO CARTESIAN GRID BOUNDARY CONDITION SPECIFIED.
AT LEAST ONE BC_TYPE MUST START WITH CG (FOR EXAMPLE CG_NSW)
RUN ABORTED

The overall box is bigger than the three boxes together - this should not be the problem.

Could somebody explain this error in more detail. I also attach the case files. geometry.stl (4.4 KB) Test_mesh.mfx (16.1 KB) BOX1_FACETS_IGNORED.stl (25 Bytes) Test_mesh.zip (50.5 KB)

#3

Hi, welcome to the MFiX forum.

Because you have defined an STL geometry, the solver wants to know what to do with it - is it a no-slip wall, free-slip-wall, etc? (The solver could just ignore the STL but that’s not how it’s implemented presently.) You need to go into the “Regions” pane, and create a new region which consists of the STL boundary (first click “+” to create a new region, then “Select facets (STL)”

shot-2021-03-18_17-09-50
shot-2021-03-18_17-09-501007Ă—967 201 KB

Once you’ve created this region, you can define a boundary condition over it.

Please let us know if this helps, or if you are still having difficulties.

– Charles

2 Likes
#4

Dear Charles,

thank you very much for your help! This was indeed the problem.
I changed the definitions of the boundary conditions and used the stls there. No the error message is gone and it seems a mesh is created.

But there is still an Error with the initial conditions.
I tried to define initial conditions in the whole bounding box → I think the error is then related to the problem that initial conditions are defined, where there is no cell/cells.
I tried to define the initial conditions only in the 3 boxes → then I have an error, because I have not defined an initial condition for each cell

Could you also give me a hint how to solve this problem?
I attached the modified case setup.
-EvaTest_mesh_adopted.zip (202.9 KB)

#5

Hi Eva,

Instead of defining the fluid domain with the three boxes (they should be unioned together to create a single volume), I flipped it around and just used a single box for the cut-out. With the normals facing outward from the box, cut-cell will mesh the correct side. I also used a point source to place the gas inlet.

image
image531Ă—632 244 KB

When I tried to run this, I am getting a segfault down in the wall function. @jeff.dietiker seems like cut-cell and k-epsilon are not working?

#6

k-epsilon should work with cutcells. Needs a non-zero value for epsilon for all ICs and MI BCs though.

#7

Thanks for the help. With your tip I managed to mesh the geometry and start the solver.

Could you specify how you defined the point source? I cannot define a point source on the stl. I used the stl to use an massinflow condition.

The solver is running, but crashing soon (ERROR time_step.f:193
DT < DT_MIN. Recovery not possible!). I guess, this is a setup problem.

Is it somehow possible to use a cyclic boundary condition for the front and back planes?