[Taghi]

Hi Foamers,
I am trying to simulate a propeller witch I created in propCad 2005 software.
There is no problem in Rhino, when importing the point cloud from propCad, But when I export it as STL and run snappyHexMesh command, the edges go crazy.
please Help me if you can.

BadMesh.png

[akidess]

You need to refine your mesh much stronger at the edges, and best also use explicit feature refinement.

[KaLium]

I had similar problems when I was using too coarse mesh. It's easy to fix.

[Taghi]

the problem is that when using a finer mesh, solution will diverge. no matter how small is the time step.

[KaLium]

Originally Posted by Taghi

the problem is that when using a finer mesh, solution will diverge. no matter how small is the time step.

What refiment levels did you use for the surfaces? What is the value of nCellsBetweenLevels?

[Taghi]

All is set on 5

[student666]

Try to check your starting surface with surfaceCheckMesh utility.
Does it pass all Checks?

SHM needs at least three cells between surface, to define properly surfaces. as already suggested maybe you have to increase refinement on edges.

[Taghi]

This is the results of surfaceCheck command:

Surface has no illegal triangles.

Triangle quality (equilateral=1, collapsed=0):
    0 .. 0.05  : 0.0273287
    0.05 .. 0.1  : 0.034305
    0.1 .. 0.15  : 0.0312111
    0.15 .. 0.2  : 0.0223543
    0.2 .. 0.25  : 0.0189269
    0.25 .. 0.3  : 0.0315751
    0.3 .. 0.35  : 0.0126786
    0.35 .. 0.4  : 0.0175923
    0.4 .. 0.45  : 0.0399769
    0.45 .. 0.5  : 0.0435257
    0.5 .. 0.55  : 0.0263884
    0.55 .. 0.6  : 0.0207164
    0.6 .. 0.65  : 0.0407049
    0.65 .. 0.7  : 0.0490157
    0.7 .. 0.75  : 0.0532015
    0.75 .. 0.8  : 0.0782857
    0.8 .. 0.85  : 0.071825
    0.85 .. 0.9  : 0.10428
    0.9 .. 0.95  : 0.13361
    0.95 .. 1  : 0.142497

    min 2.4095e-006 for triangle 17056
    max 0.999962 for triangle 15498

Edges:
    min 2.72822e-005 for edge 2343 points (-0.00077024 -0.0368108 0.000154718)(-0.000764351 -0.0368108 0.000181357)
    max 0.0503 for edge 8803 points (-0.02515 0.0068669 -1.57673e-006)(0.02515 0.0068669 -1.57673e-006)

Checking for points less than 1e-6 of bounding box ((0.176563 0.0926035 0.177137) metre) apart.
    close unconnected points 2970 (2.46516e-012 0.0068669 0.0251484) and 62 (3.30524e-012 0.0068669 0.0251484) distance:8.40085e-013
    close unconnected points 2971 (2.46516e-012 0.0068669 -0.0251516) and 1568 (3.30525e-012 0.0068669 -0.0251516) distance:8.40087e-013
Found 2 nearby points.

Surface is not closed since not all edges connected to two faces:
    connected to one face : 2113
    connected to >2 faces : 148
Conflicting face labels:2609
Dumping conflicting face labels to "problemFaces"
Paste this into the input for surfaceSubset

Number of unconnected parts : 16
Splitting surface into parts ...

Writing zoning to "zone_propellerTip.vtk"...

writing part 0 size 6494 to "propellerTip_0.obj"
writing part 1 size 129 to "propellerTip_1.obj"
writing part 2 size 2317 to "propellerTip_2.obj"
writing part 3 size 2849 to "propellerTip_3.obj"
writing part 4 size 2849 to "propellerTip_4.obj"
writing part 5 size 2317 to "propellerTip_5.obj"
writing part 6 size 129 to "propellerTip_6.obj"
writing part 7 size 2849 to "propellerTip_7.obj"
writing part 8 size 2317 to "propellerTip_8.obj"
writing part 9 size 129 to "propellerTip_9.obj"
writing part 10 size 2849 to "propellerTip_10.obj"
writing part 11 size 2317 to "propellerTip_11.obj"
writing part 12 size 129 to "propellerTip_12.obj"
writing part 13 size 2849 to "propellerTip_13.obj"
writing part 14 size 2317 to "propellerTip_14.obj"
writing part 15 size 129 to "propellerTip_15.obj"

Number of zones (connected area with consistent normal) : 17
More than one normal orientation.

[student666]

Ok, one BIG issue is your stl.
It's better if you remesh it. Try to use Salome.
Have a look to holzmanncfd.de for helping tutorials.

Sent from my ASUS_X008D using CFD Online Forum mobile app

[student666]

https://youtu.be/CZvUD3nXKa4

Sent from my ASUS_X008D using CFD Online Forum mobile app

[Taghi]

I watched the video but there it didn't have any solutions to my problem.
anyway I rebuilt the surface and rechecked it:

Region	Size
------	----
ascii	279654


Surface has no illegal triangles.

Triangle quality (equilateral=1, collapsed=0):
    0 .. 0.05  : 0.138596
    0.05 .. 0.1  : 0.149109
    0.1 .. 0.15  : 0.0963333
    0.15 .. 0.2  : 0.0455706
    0.2 .. 0.25  : 0.027838
    0.25 .. 0.3  : 0.0164131
    0.3 .. 0.35  : 0.0187339
    0.35 .. 0.4  : 0.0126978
    0.4 .. 0.45  : 0.0153011
    0.45 .. 0.5  : 0.0176504
    0.5 .. 0.55  : 0.02605
    0.55 .. 0.6  : 0.0338132
    0.6 .. 0.65  : 0.0647228
    0.65 .. 0.7  : 0.103099
    0.7 .. 0.75  : 0.0697612
    0.75 .. 0.8  : 0.0498759
    0.8 .. 0.85  : 0.0410042
    0.85 .. 0.9  : 0.0329586
    0.9 .. 0.95  : 0.0252598
    0.95 .. 1  : 0.0152117

    min 4.87625e-009 for triangle 169346
    max 0.999602 for triangle 26076

Edges:
    min 1.24006e-007 for edge 105161 points (0.0371511 0.00866685 0.0803856)(0.0371511 0.00866673 0.0803856)
    max 0.0503 for edge 357 points (2.46515e-012 0.05575 -0.0251516)(2.46515e-012 0.05575 0.0251484)

Checking for points less than 1e-6 of bounding box ((0.176562 0.0926035 0.177135) metre) apart.
    small edge between points 67691 (0.0174877 0.0539546 0.0868111) and 67690 (0.0174878 0.0539545 0.0868111) distance:2.18343e-007
    small edge between points 125663 (-0.0771587 0.0539559 0.0434571) and 125662 (-0.0771586 0.0539557 0.0434572) distance:2.22179e-007
    small edge between points 45716 (0.0879661 0.0539571 0.0101961) and 45715 (0.0879661 0.0539569 0.010196) distance:2.18811e-007
    small edge between points 102814 (-0.0651727 0.0539574 -0.0599542) and 102813 (-0.0651728 0.0539573 -0.0599541) distance:2.17539e-007
    small edge between points 102817 (-0.0651722 0.0539585 -0.0599548) and 102816 (-0.0651723 0.0539583 -0.0599547) distance:2.20831e-007
    small edge between points 102819 (-0.0651718 0.0539592 -0.0599552) and 102818 (-0.0651719 0.053959 -0.0599551) distance:2.17539e-007
    small edge between points 67698 (0.0174836 0.0539603 0.0868119) and 67697 (0.0174837 0.0539601 0.0868119) distance:2.18811e-007
    small edge between points 67699 (0.0174835 0.0539604 0.086812) and 67698 (0.0174836 0.0539603 0.0868119) distance:2.22218e-007
    small edge between points 125669 (-0.0771613 0.0539629 0.0434525) and 125668 (-0.0771612 0.0539627 0.0434526) distance:2.18049e-007
    small edge between points 102832 (-0.0651686 0.0539655 -0.0599587) and 102831 (-0.0651687 0.0539654 -0.0599586) distance:2.19381e-007
    small edge between points 102834 (-0.0651684 0.0539661 -0.0599589) and 102833 (-0.0651684 0.0539659 -0.0599589) distance:2.22147e-007
Found 11 nearby points.

Surface is closed. All edges connected to two faces.

Number of unconnected parts : 1

Number of zones (connected area with consistent normal) : 1


End

but the result is not fine yet.

[student666]

Well, I'm not so pessimistic, before you had 17 unconnected regions, now you have one region.. there are not manifolds and you solved the normal orientation.
That's an encouraging step forward.

Ok, maybe you have to change to MEfisto or Netgen 2D, the overall triangle quality is not so satisfying...

About small edges i think you're trying to mesh very fine that area. Is it strictly necessary? I think you can increase your mesh refinement phase and use the featureEdge settings for it.

 edge.emesh level (...)

Ok, one thing you have to know is that SHM needs at least 3cells between two surfaces in order to properly define both of them, during snapping step.
To get it:
Consider your initial cell dimension (background mesh) and divide by 3, that's the minimum dimension for each of the above cells that you have to obtain during castellated step.
Then, measure your small distance into your geometry and divide it by 2^n where exponent n is the refinement level you have to use to have at least the minimum cell dimension, you choose n.



Sent from my ASUS_X008D using CFD Online Forum mobile app

[student666]

In any case, consider that i face similar problem as yours recently and i was not able to solve it except for having a very big mesh size...millions and millions of cells...maybe others have other hints i don't know about...

Sent from my ASUS_X008D using CFD Online Forum mobile app

[Taghi]

Fantastic!

 edge.emesh level (...)

it worked!!!!
thank you so much.