[Nikolac]

Hi everyone,

I find an odd artefact in my simulations and I wonder if anyone has an idea what the reason for this might be:

I simulate incompressible flow in a closed 2D domain (see thumbnail below, all boundaries are impermeable walls with non-slip condition, front and back are 'empty'). The round object at the bottom of the box deforms in time, setting the surrounding fluid in motion. The deformation is such that the volume of the domain is always conserved. I essentially use the pimpleDyMFoam solver, which I modified only to implement the boundary deformation at the object (I move individual nodes of the boundary by precomputed values).

The simulation runs without any apparent problems, but in the solution I find a region next to the object where fluid would appear to flow in and out through the boundary, in contradiction to the imposed boundary condition (see the attached artefact.pdf with a picture of this). The effect is confined to a single cell of the boundary (which otherwise seems no different than any other such cell) and a small adjacent region in the interior of the domain.

I created the mesh in SALOME - checkMesh complained about unused points (for which I'm not sure why they are there), but the simulation ran anyway - so is it simply this mesh problem that comes to the surface here?
Here is the checkMesh log:

Create time

Create polyMesh for time = 0

Time = 0

Mesh stats
    points:           60120
    internal points:  20040
    faces:            136184
    internal faces:   57391
    cells:            38715
    faces per cell:   5
    boundary patches: 5
    point zones:      0
    face zones:       0
    cell zones:       0

Overall number of cells of each type:
    hexahedra:     0
    prisms:        38715
    wedges:        0
    pyramids:      0
    tet wedges:    0
    tetrahedra:    0
    polyhedra:     0

Checking topology...
    Boundary definition OK.
    Cell to face addressing OK.
 ***Unused points found in the mesh, number unused by faces: 20040 number unused by cells: 20040
  <<Writing 20040 unused points to set unusedPoints
    Upper triangular ordering OK.
    Face vertices OK.
    Number of regions: 1 (OK).

Checking patch topology for multiply connected surfaces...
    Patch               Faces    Points   Surface topology                  
    Wall                1121     2244     ok (non-closed singly connected)  
    reticularLamina     57       116      ok (non-closed singly connected)  
    HCedge              185      372      ok (non-closed singly connected)  
    back                38715    20040    ok (non-closed singly connected)  
    front               38715    20040    ok (non-closed singly connected)  

Checking geometry...
    Overall domain bounding box (-8e-05 -2.27273e-05 0) (0.0002 0.00018 1e-05)
    Mesh (non-empty, non-wedge) directions (1 1 0)
    Mesh (non-empty) directions (1 1 0)
    All edges aligned with or perpendicular to non-empty directions.
    Boundary openness (4.91637e-18 -9.05108e-18 -1.03072e-14) OK.
    Max cell openness = 2.50802e-16 OK.
    Max aspect ratio = 2.03434 OK.
    Minimum face area = 1.34262e-13. Maximum face area = 3.54092e-11.  Face area magnitudes OK.
    Min volume = 1.34262e-18. Max volume = 4.95254e-17.  Total volume = 5.00625e-13.  Cell volumes OK.
    Mesh non-orthogonality Max: 19.8124 average: 4.71872
    Non-orthogonality check OK.
    Face pyramids OK.
    Max skewness = 3.42953 OK.
    Coupled point location match (average 0) OK.

Failed 1 mesh checks.

My guess was that the problem is due to an insufficiently well conserved domain volume and thus a problem satisfying the continuity equation - however, increasing the precision in the deformation calculation, which implies more precise volume conservation of the domain, had no effect on the artefact.

If anyone has seen something like this or has a guess where it comes from (I have the feeling it must be some stupid rookie mistake on my part), any comment is highly appreciated! Let me know if you need any further information.

Regards and thanks in advance,
Nikolac