OpenFoam, SALOME, all free tools for engineers
blockMesh,snappyMesh and experimental results for 2D airfoil
Posted April 1, 2013 at 05:17 by skeptik
A certain Mahdi wrote me via e-mail, that my tetra-mesh and hybrid mesh was not good for simulations.
So here i show results which were obtained with C-mesh generated by blockMesh utility OF, and snappyHexMesh.
blockMesh domain
mesh near wall
snappyHexMesh domain
boundary layer discretization
finally, results of both mesh
The simulation's results lines are close to each other espesially when alpha<alpha critical. When alpha>alpha critical the discrepancy presents due to bad mesh at the tail in snappyHexMesh variant.
Anyway simulation results obtained from hexa mesh are similar to results were obtained from tetra mesh. So i think it doesn't make sence which mesh is used for SIMPLE calcs.
PISO results coming soon. It seems that it would be better than PISO on tetra meshes.
So here i show results which were obtained with C-mesh generated by blockMesh utility OF, and snappyHexMesh.
blockMesh domain
mesh near wall
snappyHexMesh domain
boundary layer discretization
finally, results of both mesh
The simulation's results lines are close to each other espesially when alpha<alpha critical. When alpha>alpha critical the discrepancy presents due to bad mesh at the tail in snappyHexMesh variant.
Anyway simulation results obtained from hexa mesh are similar to results were obtained from tetra mesh. So i think it doesn't make sence which mesh is used for SIMPLE calcs.
PISO results coming soon. It seems that it would be better than PISO on tetra meshes.
Total Comments 12
Comments
-
Hi skeptik!
I'm currently doing some simulations of flow over airfoils, but i have some trouble getting a nice mesh with snappyHexMesh...
could you please send me your snappyhexMeshDict you used for this case, cause the mesh looks very nice!
Thank you in advance
ArminPosted April 17, 2013 at 14:05 by arming
-
eventually it's not a good mesh. I just exceeded amount of files can be loaded to the post. The problem was described here. And i still don not know how to generate good mesh at the tail of the foil.Quote:Originally Posted by arming
Hi skeptik!
I'm currently doing some simulations of flow over airfoils, but i have some trouble getting a nice mesh with snappyHexMesh...
could you please send me your snappyhexMeshDict you used for this case, cause the mesh looks very nice!
Thank you in advance
ArminPosted April 18, 2013 at 02:23 by skeptik
-
Archive with case files is here
https://www.dropbox.com/s/1ih182x44v...appyHex.tar.gz
There are both snappyHex case and 2Dextrusion case with extrusion dictionary. In the one of this case resulted mesh was saved and archived.
Enjoy.Posted April 18, 2013 at 04:13 by skeptik
-
Hi skeptik!
I'm currently trying to simulate the flow over a high aspect ratio ellipse but i am very new with snappyHexMesh and blockMesh. I always imported my meshes from gambit...
would you mind to repost the snappyhexMeshDict and blockMeshDict you used for this case? The link is no longer available...
Thanks in advance!
Best!Posted April 24, 2014 at 09:24 by maalan
-
Now i have no sHM files on my computer.. May be tomorrow i'll find it.Quote:Originally Posted by maalan
Hi skeptik!
I'm currently trying to simulate the flow over a high aspect ratio ellipse but i am very new with snappyHexMesh and blockMesh. I always imported my meshes from gambit...
would you mind to repost the snappyhexMeshDict and blockMeshDict you used for this case? The link is no longer available...
Thanks in advance!
Best!
But, ellipse is an analytical surface. You really don't need a sHM utility to build a mesh. Try blockMesh!Posted April 24, 2014 at 13:27 by skeptik
-
Thanks for your reply, skeptik!! I know what you mean, and up to the moment I tried successfully a conformal mesh from a cylinder mesh (I applied the Joukowski transformation). The point is that my ellipse has 2% aspect ratio, so, assuming the major axis of my ellipse is aligned with the x-axis and the minor axis with the y-axis, i get very thin cells normal to the wall both in the trailing and leading edges and much coarser cells normal to the wall in the middle of my ellips. I'm afraid that my viscous term will not be well solved... What do you think of?? any idea??
Best!Posted April 24, 2014 at 13:56 by maalan
-
what is your blocking structure?Quote:Originally Posted by maalan
Thanks for your reply, skeptik!! I know what you mean, and up to the moment I tried successfully a conformal mesh from a cylinder mesh (I applied the Joukowski transformation). The point is that my ellipse has 2% aspect ratio, so, assuming the major axis of my ellipse is aligned with the x-axis and the minor axis with the y-axis, i get very thin cells normal to the wall both in the trailing and leading edges and much coarser cells normal to the wall in the middle of my ellips. I'm afraid that my viscous term will not be well solved... What do you think of?? any idea??
Best!Posted April 24, 2014 at 14:19 by skeptik
-
My original mesh is a circular cylinder within an O-grid domain. Once applied the Joukowski transformation I get an ellipse within an O-grid conformal mesh. Have a look here, skeptik!!
https://www.dropbox.com/s/lp2mbo0sz2jblgt/ellipse.pdf
https://www.dropbox.com/s/jgmjbbv9mlvormu/ellipse2.pdf
https://www.dropbox.com/s/owr757rkbwi386m/ellipse3.pdfPosted April 24, 2014 at 14:41 by maalan
-
You're really smart guy, cause you know about Jukowsky transformation. But i think this way too complicated.Quote:Originally Posted by maalan
My original mesh is a circular cylinder within an O-grid domain. Once applied the Joukowski transformation I get an ellipse within an O-grid conformal mesh. Have a look here, skeptik!!
https://www.dropbox.com/s/lp2mbo0sz2jblgt/ellipse.pdf
https://www.dropbox.com/s/jgmjbbv9mlvormu/ellipse2.pdf
https://www.dropbox.com/s/owr757rkbwi386m/ellipse3.pdf
Just make the o-grid blocks with elliptical edges with attention to edge-to-egde distance and you will have a good mesh to resolve BL.
https://www.dropbox.com/s/slsu1ek052z011c/ellipse.png
Posted April 25, 2014 at 07:29 by skeptik
-
Thank you skeptik! I considered your idea but my ellipse has an aspect ratio equal to 2% (try to plot it and you will understand me), so this is the reason why I decided emulate your O-grid and mesh it by using sHM and adding some layers... So, could you share again your files, please??
Thank you very much!!
All the best.Posted April 25, 2014 at 08:36 by maalan
-
No matter how big your AR is. This way to build meshes with o-grid block works fine also in case of thin compressor blades.Quote:Originally Posted by maalan
Thank you skeptik! I considered your idea but my ellipse has an aspect ratio equal to 2% (try to plot it and you will understand me), so this is the reason why I decided emulate your O-grid and mesh it by using sHM and adding some layers... So, could you share again your files, please??
Thank you very much!!
All the best.
I will look for sHM - files. I just wanna to say reduce nLayers and increase minMedianAxisAngle and featureAngle. Your empty patches should be "patches" when you generating coarse and snapped mesh. You should switch to "empty" only when generating layers.
Any other settings you could see here http://openfoamwiki.net/images/f/f0/...SlidesOFW7.pdfPosted April 25, 2014 at 08:56 by skeptik
-
Hi skeptik, I approach your same problem with snappyHexMesh but your mesh seem better than mine!! you can repost the link to your snapyHexMeshDict please. Thank for your attention.Quote:Originally Posted by skeptik
No matter how big your AR is. This way to build meshes with o-grid block works fine also in case of thin compressor blades.
I will look for sHM - files. I just wanna to say reduce nLayers and increase minMedianAxisAngle and featureAngle. Your empty patches should be "patches" when you generating coarse and snapped mesh. You should switch to "empty" only when generating layers.
Any other settings you could see here http://openfoamwiki.net/images/f/f0/...SlidesOFW7.pdf
Posted December 10, 2020 at 04:40 by aglios900
