|
distorted o-grid and bad cells quality
4 Attachment(s)
Hello can anyone please help me,
I am trying to mesh a pipe geometry (pic 1), which consists of two elbows (lower and upper, S shape pipe, ) in Ansys Icem. Problem_1 --> For better quality i had created an O-grid which resulted in distorted edges in a zig zag shape in both elbows(upper and lower elbows, can be seen in the pictures i.e pic 2 & 3). Please suggest me anything for it. Problem_2--> After fixing it manually using moving vertices, the quality somehow improved but cannot be improved further due to bad elements lying in the geometry (shown in pic 4). It would be really generous and kind of you if some one can suggest me anything Please. I have been trying for one week and tried everything but cannot improve the quality. Since i have to do grid independence study on this geometry. PLEASE help me out :( |
Did you try to align the interior vertices (light blue) with the outer vertices (white)?
Blocking>Move Vertex>Align Vertices in-line |
yes i aligned these edges (blue lines with the white ones) after words, but the problem is, even after aligning, the cells lying on the blue lines (connecting two faces i.e back and front) are having bad quality :(
|
The cells shown in the last picture
|
You don't show elements, you only show blocks. So difficult to say.
You show a histogram with quality. It looks like you marked elements with quality around 0.6 to have bad quality. I don't consider that as bad quality. What do you consider as bad quality? Did you perfrom a test that proved that 0.6 is bad? |
No actually i want to have the cells minimum quality 0.7 As i would be doing grid independence study later on in ansys fluent. Just for the better results and convergence
|
Ok. But I still don't see elements. I see mainly blocks, and some bunching over the length. But how many elements did you specify in radial and tangential direction?
|
Sorry, i miss understood about blocks, here is the snapshot showing blocks https://drive.google.com/file/d/1aqZ...ew?usp=sharing
Actually i didnt specify the number of nodes by myself, i used Mesh--> Global Mesh setup---> Max.element size. This is what i have done to get the mesh, since i would be doing grid independence study for different number of cells. So what i thought that i would be changing the max element size and will be generating the finer and finer mesh. Please correct me if needed . |
I don't want to see blocks. I want to see your elements.
So, hide the blocking, show elements....... |
Elements snapshot https://drive.google.com/file/d/1Ys-...ew?usp=sharing
Can you please have a look on the project as well?? so that you can have a better idea https://drive.google.com/drive/folde...Q7?usp=sharing |
Why don't you show the Pre-mesh? Or the mesh after conversion to unstructured mesh?
|
Sorry Gert, you can see that in the .prj file that i have shared. I mean you can just have a look by yourself. The screen shots quality is not coming good for the mesh
|
|
Mesh looks fine to me. You could opt for a growth factor before and after the inner bend to reduce the jump in cell size
Alternatively, in the inner bend you could move the blocking a bit upstream and downstream. But you get skewed elements in return. |
Thank you so much Gert-Jan, it worked :) . I really appreciate that :)
|
| All times are GMT -4. The time now is 13:36. |