face normal direction of a wedge
 

http://s30.postimg.org/fgcq0ylm9/Untitled.pnghttp://postimg.org/image/5j1p7we0d/

I tried to create a wedge for a axial symmetrical problem. This figure is the geometry. Following is my
blockMeshDict file.

According to the user guide, the normal direction of any boundary face should point to outside of the computational zone. So I defined all the faces in this way. Error messages popped out when I run blockMesh, saying outlet and axi_symm-r don't have neighbor cell faces.

I found that if I change the vertices in the reversed order (the quoted lines) I can generate the mesh successfully. But in this case, the normal directions of these two faces are actually pointing into the computational zone. I wonder what is wrong. Thanks

convertToMeters 1;

vertices
(
(0 0 0) //0
(8 0 0) //1
(0 -0.0034905 0.099985) //2
(8 -0.0034905 0.099985) //3
(8 0.0034905 0.099985) //4
(0 0.0034905 0.099985) //5
);

blocks
(
hex (0 1 1 0 2 3 4 5) (100 1 20)
simpleGrading
(
1
1
(
(0.5 0.5 4)
(0.5 0.5 0.25)
)
)

);

edges
(
arc 2 5 (0 0 0.1)
arc 3 4 (8 0 0.1)
);

patches
(
patch
inlet
(
(0 2 5 0)
)

patch
outlet
(
// (1 3 4 1)
(1 4 3 1)
)

wedge
axi_symm-f
(
(0 1 3 2)
)

wedge
axi_symm-r
(
// (0 1 4 5)
(0 5 4 1)
)

wall
wall
(
(2 3 4 5)
)
);

mergePatchPairs
(
);

// ************************************************** *********************** //

Hi popo, or kaka or whatever (I hope you aren't german :D )

You made an fault by defining your block.
If you want to let collapse some points to create a prism... don't let the first 4 collapse, as they are used to define your local coordinate system of the block.

I hope it helps
regards
Alex

Thanks for the response.

I changed the geometry. So now it is like:
http://s27.postimg.org/c7l9rgo77/Untitled.png

However, I still had the same error message:
--> FOAM FATAL ERROR:
face 0 in patch 0 does not have neighbour cell face: 4(4 0 1 4)

I didn't find it in the user guide. But does the symmetrical axis need to be one of the coordinate axis?

thanks

hmmm, looks good so far...

but I am not sure about your multi-grading-thing... never done that... I hope you are sure that there is no error.

Furthermore I always rather describe the boundary than the patches.
like here http://cfd.direct/openfoam/user-guide/blockmesh/ see under 5.3.1.5 The boundary

I hope it helps

Edit:
Another possible error could be that you haven't defined the empty patch (4 5 5 4) . But I don't know it refers to the error message.

I found this example:
https://openfoamwiki.net/index.php/M...s/AxiSymmetric

But even in this case, the normal direction of the outlet is pointing to the computational domain. I am really confused...

the rule to set the boundary is defined in the users guide as
but I thought it fits to your dict

yeah, in my case, it fits. But the case doesn't work.

All the working cases I saw so far have one or two faces violating this rule. I am really confused...

sorry, but I am at my wits end.

Have you defined the empty face?

I tried that as well.

Thanks for your help anyway

there is no try... do it ;)

I wanted to say that it's necessary to define the collapsed face explicitly. (as I read it in the guide)

And one thing I wanted to say at first when I saw your dict (but I forgot it) :

I think you have to write:
Instead of:

I did.

Take a look at this example:
https://openfoamwiki.net/index.php/M...s/AxiSymmetric

The normal direction of the outlet is in the "wrong direction". If you change that, the case won't work. Everything else you mentioned was in this case. The empty face. The way of defining faces for the new version of openFOAM