checkMesh for a Hollow Cylinder
2 Attachment(s)
ok I am rather new to CFD and to OF as well. So please kindly be patient and forgiving.
I am now attempting a simple flow analysis to get myself started. And I have a problem understand where I have checkMesh errors. The case is a simple hollow cylinder. I need to create a mesh to represent the hollow cylinder. I used blockMesh to do this. Each block must be 8 vertices ; i have a hollow part ; so I need to have to have at least 2 blocks. When I was doing this, I used 4 blocks. These 4 blocks were in the form of 4 quadrants starting anticlock from 3 o'clock (see figure Attachment 35054). I defined each of the 4 blocks starting from the inner point outwards. Since it is a enclosed cylinder, I have merge the facing pairs of each quadrant using mergePatchPairs in blockMeshDict. And this I what I have previewed using paraview. Looks ok. However, checkMesh tells me I have negative volume and I do not understand since the preview in paraview looks ok. Attachment 35055 That being said, running the solver also does not work, as it crashed out immediately. I assume it is the problem with the mesh. Here is the checkMesh message. Code:
Mesh stats |
Whenever you get checkMesh errors you can see where the errors are by checking the directory sets that is created within the case directory after runing checkMesh. You can see these errors in paraview so that is easier to find where you did something wrong. However I guess that the problem is wrong orientation of the blocks or patches in the blockMeshDict file....
Regards, Alex |
Quote:
In the definition of the Blocks e.g. (0 1 2 3 4 5 6 7), must 1 and 2 be defined in the positive x1 and x2 direction or can it be negative x1 and x2 direction? Im suspecting this could be the problem. I couldnt find the answer to my question in the manual. so thats why I am asking here. |
It's not a matter of defining all the points in the positive direction of the axes, but it's a matter of the point order in the block/patch definition. Maybe you haven't paid enough attention to this section of the manual...
|
Quote:
Im asking this question cos I realised defining my 2nd quadrant in the negative x1 and x2 direction gives a different error message. However, now I have rebuilt my BlockMeshDict using just 2 blocks - top and bottom half. Also, I used the same defined vertices to define both block for the common points. so I did not have to define common faces in mergePatchPairs. Weirdly, there are still error messages in checkMesh, but only 1 Mesh has failed, which I cannot understand why. Code:
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * // |
You can check the fail with paraview opening the set nonAlignedEdges. Why don't you upload your blockMeshDict file so that someone can check your process?
|
Quote:
Code:
/*--------------------------------*- C++ -*----------------------------------*\ |
The error shown by checkMesh only means that you forgot to define the boundary faces in your blockMeshDict file. By the way, two questions.
First one, you forgot to define one of the arcs, well actually you have defined it in a wrong way. This is the arc going from point 15 to point 8 that you have defined it from 15 to 0. Second question, why are you using such an amount of blocks to create a cylinder? You can do it with less blocks. Look at your blockMeshDict modified to create your cylinder with the half of the blocks you used: Code:
/*--------------------------------*- C++ -*----------------------------------*\ Regards, Alex |
Quote:
|
1 Attachment(s)
Hello :)
based on this link (https://www.ehsanmadadi.com/cylinder-mesh/) I created a hollow cylinder as .m4 file for my project. Maybe someone can use it for other tasks. To use the file you only have to change the extension .txt to .m4. Best regards Kevin |
All times are GMT -4. The time now is 11:49. |