[snappyHexMesh] Question about mesh produced by snappyHexMesh

bigphil · May 14, 2009, 07:49

Hi,

I only started using snappyHexMesh recently, and it seems like a really nice utility.

One problem: I thought that all nodes in a mesh must join ie one cannot have a node in the middle of a face. But the mesh produced by snappyHexMesh, shows mesh edges ending the the middle of other edges, see attached picture from OpenFOAM manual about snappyHexMesh. The red circle showing one of many.

Surely, this cannot be right? Maybe I am missing something?

Could someone please explain?

bigphil · May 14, 2009, 12:04

One other quick question:

What is the 'cellLevel', that snappyHexMesh prints out?

Is this something to do with the type of cells or quality of them?

Any enlightening is welcome,

Philip

wolle1982 · May 15, 2009, 05:20

Hi,

don't worry, OF creates an addaption layer between the refinementlevels. The picture is just to explain, it doesn't show the real meshing process of sHM.

But here's another question to you: why are you sure it is a hex-block with a node too much? Could this also be a polyhedral Block with 10 nodes....?!

bigphil · May 15, 2009, 05:38

Quote:

Originally Posted by wolle1982

But here's another question to you: why are you sure it is a hex-block with a node too much? Could this also be a polyhedral Block with 10 nodes....?!

aaahhh I see! That makes sense.

Thanks very much for the reply, as I was quite confused by this.

Philip

fra76 · May 15, 2009, 06:37

Hi Phil,
as Wolfgang pointed out, the transition cells between two different refinement levels are polyhedrons (the big one above the red point in your picture, for example).
The cellLevel volScalarField saved during the meshing process is the refinement level. If you open the mesh in paraFoam and colour the volume by that variable, you will see that all the cells on the same refinement level (i.e. with the same size, apart deformation induced by snapping) have the same colour.

Hope this helps,
Francesco

bigphil · May 15, 2009, 06:49

Francesco,

Thank you, that does help. Thanks for clearing that up.

snappyHexMesh is a really nice tool, and I am trying to understand it as best I can.

One other question:
When one uses 'addLayers' to add layers of cells to a surface patch, does snappyHexMesh try to add all hex cells, or mostly hex with some polyhedra, or just polyhedra?

I thought it was meant to add a layer of hex, but I am trying out snappyHexMesh on a ball, and it seems to add all polyhedra. Although I am not sure about the settings I am using in the 'addLayers' section in the dict.

fra76 · May 15, 2009, 07:01

Hi Phil.
The layers are added after the snapping phase. The shape depends on the shape of the elements on the surface mesh (mostly quads, hence hexa elements as prisms). Be aware, however, that for quality issues the surface elements can sometimes be merged together, creating polyhedrons when the prisms are build on top of them.

Francesco

wendywu · June 30, 2009, 22:31

Hi,

I am confused by my problem.
I am trying to mesh on a flow field in screw extruder.
that means a screw extruder inside a barrel pushing plastic material to moving from one side to another.
I made a screw and created the stl.
I blockmeshed a cylindrical background(barrel).
then I began to sHM. in snappyHexMeshDict, I set locationInMesh just inside the flow field(between the barrel and the screw).
but it always exited with information that point is not inside mesh or on a face or edge. I did tried avoiding choosing a point on a cell face or edge. and I am sure the point is inside mesh.
But I can not figure out why it always exit.
anybody can give me some guidance?
Thank you very much.
Wendy

wolle1982 · July 1, 2009, 10:27

hi wendywu

i also had that problem once. i think the solution was to make sure the coordinatesystem of both, model and blockmesh-mesh have to be the same.

also watch the units. a 1m solid inside a 10mm mesh is not cooking.

wendywu · July 1, 2009, 13:11

Thank you. Wolfgung.
I have solved the problem, as you mentioned. I changed the vertexes.
But when it run sHM for some iterations, I don't know it is finished or not. it says : killed.(what does this mean?) then stoped.
I didn't see any SHM results in paraFoam or in documents.
Thank you again. I can not fingure out what is the problem.

Wendy

J-P · March 25, 2010, 08:27

I'm new with SnappyHexMesh (and OpenFOAM too) and trying figure it out. Right now I'm just testing the cell splitting feature i.e. I'm hoping to start with a mesh of big cubes and end up with a mesh with smaller ones so that the angles stay as right-angles.

I've attached a result I got and I'm a bit worried about the cells between the two refinement levels as their angles don't seem to be right anymore. However, I've taken a look at the files "points", "faces" and "owner" which seem to tell me that the cells that look split in paraFoam actually aren't split.

I've heard about paraFoam's problem to display polyhedra, do you think this might be the case here too?

The cell I examined more closely owned total 9 faces so that these faces formed a cube with one side open and so that 4 of the OF-faces made up one face of the cube. See the other attached picture to get a better idea of what I'm talking about. Of the mentioned faces, all but the one at the bottom were in the "owner" file. My interpretation of this is that there is no need to explicitly declare the bottom face as this is required to make the cube "complete". Is this so or am I missing some point here?

Thanks for all replies!

wendywu · March 26, 2010, 20:35

1) end up with a mesh with smaller ones so that the angles stay as right-angles.

I am not sure about this: It is not guaranteed that the right angles would always be right angles in the mesh.

2) I've heard about paraFoam's problem to display polyhedra, do you think this might be the case here too?

I don't know the problem, but I don't think the result is caused by the display problem.

3) The cell I examined more closely owned total 9 faces so that these faces formed a cube with one side open and so that 4 of the OF-faces made up one face of the cube.
one side open means there is no face in the side?

4) Of the mentioned faces, all but the one at the bottom were in the "owner" file. My interpretation of this is that there is no need to explicitly declare the bottom face as this is required to make the cube "complete". Is this so or am I missing some point here?

I think if you want to do the simulation, there should be a boundary even though it is a open side, so that you can apply a boundary condition to it, and the calculation area is "complete" . Do you think so?

I am not sure if my understanding is correct or not, just give you some reference idea.

Wendy

Thanks for all replies![/QUOTE]

June 30, 2009, 22:31	location in mesh when sHM	#8
wendywu Member xianghong wu Join Date: Mar 2009 Posts: 57 Rep Power: 17	Hi, I am confused by my problem. I am trying to mesh on a flow field in screw extruder. that means a screw extruder inside a barrel pushing plastic material to moving from one side to another. I made a screw and created the stl. I blockmeshed a cylindrical background(barrel). then I began to sHM. in snappyHexMeshDict, I set locationInMesh just inside the flow field(between the barrel and the screw). but it always exited with information that point is not inside mesh or on a face or edge. I did tried avoiding choosing a point on a cell face or edge. and I am sure the point is inside mesh. But I can not figure out why it always exit. anybody can give me some guidance? Thank you very much. Wendy

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
SnappyHexMesh... Domain boundaries are uneven	Blue8655	OpenFOAM Pre-Processing	0	January 20, 2016 11:43
[mesh manipulation] Importing Multiple Meshes	thomasnwalshiii	OpenFOAM Meshing & Mesh Conversion	18	December 19, 2015 18:57
[OpenFOAM] snappyHexMesh - can't open mesh in ParaView	tfigueiro	ParaView	0	June 11, 2015 19:05
fluent add additional zones for the mesh file	SSL	FLUENT	2	January 26, 2008 11:55
How to control Minximum mesh space?	hung	FLUENT	7	April 18, 2005 09:38

May 14, 2009, 12:04		#2
bigphil Super Moderator Philip Cardiff Join Date: Mar 2009 Location: Dublin, Ireland Posts: 1,089 Rep Power: 34	One other quick question: What is the 'cellLevel', that snappyHexMesh prints out? Is this something to do with the type of cells or quality of them? Any enlightening is welcome, Philip

May 15, 2009, 05:20		#3
wolle1982 Senior Member Wolfgang Heydlauff Join Date: Mar 2009 Location: Germany Posts: 136 Rep Power: 21	Hi, don't worry, OF creates an addaption layer between the refinementlevels. The picture is just to explain, it doesn't show the real meshing process of sHM. But here's another question to you: why are you sure it is a hex-block with a node too much? Could this also be a polyhedral Block with 10 nodes....?!

May 15, 2009, 06:37		#5
fra76 Senior Member Francesco Del Citto Join Date: Mar 2009 Location: Zürich Area, Switzerland Posts: 237 Rep Power: 18	Hi Phil, as Wolfgang pointed out, the transition cells between two different refinement levels are polyhedrons (the big one above the red point in your picture, for example). The cellLevel volScalarField saved during the meshing process is the refinement level. If you open the mesh in paraFoam and colour the volume by that variable, you will see that all the cells on the same refinement level (i.e. with the same size, apart deformation induced by snapping) have the same colour. Hope this helps, Francesco

May 15, 2009, 06:49		#6
bigphil Super Moderator Philip Cardiff Join Date: Mar 2009 Location: Dublin, Ireland Posts: 1,089 Rep Power: 34	Francesco, Thank you, that does help. Thanks for clearing that up. snappyHexMesh is a really nice tool, and I am trying to understand it as best I can. One other question: When one uses 'addLayers' to add layers of cells to a surface patch, does snappyHexMesh try to add all hex cells, or mostly hex with some polyhedra, or just polyhedra? I thought it was meant to add a layer of hex, but I am trying out snappyHexMesh on a ball, and it seems to add all polyhedra. Although I am not sure about the settings I am using in the 'addLayers' section in the dict.

May 15, 2009, 07:01		#7
fra76 Senior Member Francesco Del Citto Join Date: Mar 2009 Location: Zürich Area, Switzerland Posts: 237 Rep Power: 18	Hi Phil. The layers are added after the snapping phase. The shape depends on the shape of the elements on the surface mesh (mostly quads, hence hexa elements as prisms). Be aware, however, that for quality issues the surface elements can sometimes be merged together, creating polyhedrons when the prisms are build on top of them. Francesco

July 1, 2009, 10:27		#9
wolle1982 Senior Member Wolfgang Heydlauff Join Date: Mar 2009 Location: Germany Posts: 136 Rep Power: 21	hi wendywu i also had that problem once. i think the solution was to make sure the coordinatesystem of both, model and blockmesh-mesh have to be the same. also watch the units. a 1m solid inside a 10mm mesh is not cooking.

July 1, 2009, 13:11		#10
wendywu Member xianghong wu Join Date: Mar 2009 Posts: 57 Rep Power: 17	Thank you. Wolfgung. I have solved the problem, as you mentioned. I changed the vertexes. But when it run sHM for some iterations, I don't know it is finished or not. it says : killed.(what does this mean?) then stoped. I didn't see any SHM results in paraFoam or in documents. Thank you again. I can not fingure out what is the problem. Wendy

March 26, 2010, 20:35		#12
wendywu Member xianghong wu Join Date: Mar 2009 Posts: 57 Rep Power: 17	1) end up with a mesh with smaller ones so that the angles stay as right-angles. I am not sure about this: It is not guaranteed that the right angles would always be right angles in the mesh. 2) I've heard about paraFoam's problem to display polyhedra, do you think this might be the case here too? I don't know the problem, but I don't think the result is caused by the display problem. 3) The cell I examined more closely owned total 9 faces so that these faces formed a cube with one side open and so that 4 of the OF-faces made up one face of the cube. one side open means there is no face in the side? 4) Of the mentioned faces, all but the one at the bottom were in the "owner" file. My interpretation of this is that there is no need to explicitly declare the bottom face as this is required to make the cube "complete". Is this so or am I missing some point here? I think if you want to do the simulation, there should be a boundary even though it is a open side, so that you can apply a boundary condition to it, and the calculation area is "complete" . Do you think so? I am not sure if my understanding is correct or not, just give you some reference idea. Wendy Thanks for all replies![/QUOTE]