|
[Sponsors] |
![]() |
![]() |
#21 |
New Member
Deepak Cheekati
Join Date: Jul 2012
Location: Chennai, India
Posts: 21
Rep Power: 14 ![]() |
Hi Bruno
![]() Thanks for the link. I tried various geometries and tried to vary the wall thickness and all, but found that the problem is geometry independent. I resolved this problem by altering the grading of background mesh. ![]() I made sure that the background mesh of big cylinder doesnot intersect my inlet pipe. Please look at the images attached. ![]() Thank You everyone for your valuable suggestions ![]() Turbulence. |
|
![]() |
![]() |
![]() |
![]() |
#22 |
New Member
Deepak Cheekati
Join Date: Jul 2012
Location: Chennai, India
Posts: 21
Rep Power: 14 ![]() |
Thank YOU
![]() |
|
![]() |
![]() |
![]() |
![]() |
#23 | |
Member
Join Date: Jul 2012
Posts: 66
Rep Power: 14 ![]() |
Quote:
Hello Bruno, I am trying to get SHM snapping to my STL surface (see attachments). I have tried a lot and read a lot of solutions.. I have tried to play around with snapping parameters ... I increased snapping iterations (5000 iterations) ... I have disabled all snapping controls ... tried feature edge handling ... createBuffles command ... etc. I have learn from this thread (http://www.cfd-online.com/Forums/ope...tml#post383012)... and it improved a bit (see attached photo) .. but still poor snapping Same message: Did not successfully snap mesh. Giving up... It would be really nice if you took a look and give me your advice Attached are Photo of STL, photo of the snapped mesh, and the triSurface folder. from here you can download the case: http://dl.dropbox.com/u/32336583/Case.zip Really appreciate your guidance! Kind Regards, |
||
![]() |
![]() |
![]() |
![]() |
#24 |
Member
Join Date: Jul 2012
Posts: 66
Rep Power: 14 ![]() |
a little detail on how Umb patch is generated in my case
I got Umb patch from an STL file... this is what I did 1-Extracted the geometry from the STL and convert it to emesh using the command: surfaceFeatureExtract -includedAngle 150 -writeObj constant/triSurface/Umb.stl Umb 2-Duplicated edges in emesh (simply by copy-pasting) as recommended here (http://www.cfd-online.com/Forums/ope...rane-edge.html) 3-Converted this emesh file to faceZones by adding faceZone Umb; cellZone Umb; cellZoneInside inside; to snappyHexMeshDict as follows: refinementSurfaces { Umb { // Surface-wise min and max refinement level level (2 2); faceZone Umb; cellZone Umb; cellZoneInside inside; } } 4- After the SHM, I converted this faceZones into Patches using the command: createBaffles Umb '(Umb Umb)' Thanks in advance for your advice |
|
![]() |
![]() |
![]() |
![]() |
#25 | |
Member
Join Date: Jul 2012
Posts: 66
Rep Power: 14 ![]() |
Quote:
I got an answer here: http://www.cfd-online.com/Forums/ope...e-failure.html Thanks Forum! ![]() |
||
![]() |
![]() |
![]() |
![]() |
#26 |
New Member
Stefan
Join Date: Jan 2011
Location: Bremen
Posts: 20
Rep Power: 15 ![]() |
Hello!
I am having a problem with the snapping of nodes onto a cylinder surface when using snappyHexMesh. I tried all the hints in this threat, which do not help in my case. I further tried closed cylinders in a fully surrounding blockMesh, from which I cut the part I am interested in: Nothing works. I am really frustrated. can someone help me, please? I attach the case including bash script, which should illustrate the method I am trying to apply. I further attach some mesh scenes: - refined blockMesh with cylinder surface (stl file), - the castellated mesh with cylinder surface and - the snapped mesh with cylinder surface. I tried to modify any of the parameters. So far it seems to me, that using explicit feature edge refinement does not help and regionwise refinement eitern. Further, increasing iterations in the snapControls section and decreasing tolerance does not improve the result eiter. I have no idea what to try. Any advice is highly appreciated! Best regards Stefan |
|
![]() |
![]() |
![]() |
![]() |
#27 |
Member
David GISEN
Join Date: Jul 2009
Location: Germany
Posts: 68
Rep Power: 17 ![]() |
Hi Stefan,
try it again using feature edge refinement and check the .eMesh-File with paraview beforehands. Increase the level of feature refinement. I had a similar problem but fixed it with this. The main difference is that I use the edge of a closing area on top of the cylinder, not of the cylinder itself, maybe it's important. I guess it's on purpose that feature edge is commented and the .eMesh is zipped in your uploaded files? Regards, David |
|
![]() |
![]() |
![]() |
![]() |
#28 |
New Member
Stefan
Join Date: Jan 2011
Location: Bremen
Posts: 20
Rep Power: 15 ![]() |
Hi David,
thank you very much for your reply. I tried to follow your recommendations. First I must admit, that I could not load constant/trisurface/cylinder.emesh (not the packed and neither the unpacked one). At least I dont know, which filter to use. I could load cylinder_edgeMesh.obj, cylinder_openEdges.obj and cylinder_regionEdges.obj. I further extracted the coordinates from the eMesh file. All looks good to me (see 20121111_4cylinder.eMesh_points.png). The points lie on the edges and their number corresponds to the blockMesh resolution (see 20121111_4.png). I refined the blockMesh a little further. I think it is ok to be able to represent the geometry (it is actually a bit finer than in the flange case, mentioned previously in this threat). The *.eMesh file has been gzipped because I set writeCompression compressed in the system/controlDict. As recommended in several tutorials, I extended the cylinder length so that it pierces out of the blockMesh in order to improve the upper and lower edges. That's why, I don't have closing areas for the cylinder. Following your recommendation, I tried to use a cylinders which has been closed with new geometry at teh top and at the bottom. The grid looks a bit different there, but it does not solve the problem (I don't attach the figure). :-( Best regards Stefan Last edited by holodeck10; November 11, 2012 at 21:34. |
|
![]() |
![]() |
![]() |
![]() |
#29 | |
Member
David GISEN
Join Date: Jul 2009
Location: Germany
Posts: 68
Rep Power: 17 ![]() |
Hi Stefan,
you're right, the .eMesh cant be opened, but the .obj-Files should do as well to check whether the creation of the edges was successful. I really don't know why it doesnt work for you, sorry. Last thing you could try is Quote:
![]() Regards, David |
||
![]() |
![]() |
![]() |
![]() |
#30 |
New Member
Stefan
Join Date: Jan 2011
Location: Bremen
Posts: 20
Rep Power: 15 ![]() |
I found a solution, finally. Two things should be changed in my case:
1) The stl surface should surround the numerical domain instead of the cylinder to cut out. 2) Unfortunatelly, refineMesh cannot be applied in zones where it touches the stl file (see attached figure). This would lead to an error message. I try to find a workaround for beeing still able to apply refineMesh in order to change the aspect ratio of cells, e.g. by generating a larger mesh and cutting the interesting part out afterwards. Hope this is of use for you as well. Thanks to all who dedicated their time to my problem! Cheers Stefan |
|
![]() |
![]() |
![]() |
![]() |
#31 |
Member
Claudio
Join Date: Mar 2012
Location: Milano, Italy
Posts: 49
Rep Power: 14 ![]() |
Dear Foamers,
I opened a new thread about a similar problem with snappy, here: http://www.cfd-online.com/Forums/ope...tml#post445745 As you can read, I tried many kind of combinations (in the blockMeshDict and in the sHMDict, expecially in refinementSurface and snapControls), but it seems that I'm not able to do the snap stage ![]() So, if someone of you could help me (here or on my new thread) I would be very happy. I looked around in the forum, for a solution, but till now nothing seems working! Many thanks in advance! Best regards, Claudio |
|
![]() |
![]() |
![]() |
![]() |
#32 | |
New Member
Chen Yu
Join Date: Sep 2013
Posts: 19
Rep Power: 12 ![]() |
I have similar problem, snap doesn't work. The attachment shows the mesh generated by sHM (first figure) and stl file (second). There is step mesh on one cylinder, but there is no problem with another. Anyone got idea to fix the problem? Case folder and Allrun script is also attached. I'm using OF2.1.1
sHM sHM.jpg stl stl_file.jpg cylinder_snappyHexMesh.tar.gz problem solved now, mistakes in blockMeshDict where boundary should not be empty in one direction. Otherwise, snappyHexMesh will treat it as a 2D case. Also see this post http://www.cfd-online.com/Forums/ope...ped-cells.html Quote:
Last edited by chery1986; March 18, 2014 at 20:18. |
||
![]() |
![]() |
![]() |
![]() |
#33 |
Member
Aldo Iannetti
Join Date: Feb 2010
Posts: 48
Rep Power: 16 ![]() |
Hi guys
Please see the attached picture. Could you please tell me why on earth in my simple geometry I have good quality surfaces along with surfaces where the snap clearly fails? Any suggestion on which parameter in snappyHexMeshDict to adjust to fix the problem? Cheers |
|
![]() |
![]() |
![]() |
![]() |
#34 |
Retired Super Moderator
Bruno Santos
Join Date: Mar 2009
Location: Lisbon, Portugal
Posts: 10,978
Blog Entries: 45
Rep Power: 128 ![]() ![]() ![]() ![]() ![]() ![]() |
Greetings Aldo,
![]() ![]() ![]() Since I have to guess, I would say that the maximum local and global number of cells are not making sense, which leads snappyHexMesh to stop attempting at refining and cutting cells, which would increase the cell count, therefore it stops with the last known sane-looking mesh for the requested maximum cell count. Best regards, Bruno
__________________
|
|
![]() |
![]() |
![]() |
![]() |
#35 | ||
New Member
John Smith
Join Date: Jun 2015
Posts: 4
Rep Power: 11 ![]() |
Hi All,
New to OpenFoam! Ive been trying to run the turbineSiting scenario on different locaiton. I managed to include a separate STL file for the terrain and this looks well mashed. However I am finding it nearly impossible to enter a searchable box. codes are: Quote:
Quote:
|
|||
![]() |
![]() |
![]() |
![]() |
#36 |
Senior Member
Saideep
Join Date: Apr 2015
Location: INDIA
Posts: 203
Rep Power: 12 ![]() |
Hi Mark;
Welcome to the forum. As I read your mesh is generate well. Then as you use the snappyHexMesh utility you generate a new time step. Else if you use command snappyHexMesh -overwrite you generate your required shape in the same 0 time step. Now as I see your code you are interested in looking for a searchableBox which is the name of your .stl file for openFoam to know. So, once you open paraFoam and apply your case, you get all the available boundaries as mentioned in your blockMeshDict file along with the name of the boundary given for your .stl file. Hope this helps; Saideep |
|
![]() |
![]() |
![]() |
![]() |
#37 | |
New Member
John Smith
Join Date: Jun 2015
Posts: 4
Rep Power: 11 ![]() |
hi Saideep,
The mesh is indeed generating around my terrain STL. However no mesh is being drawn around my refinement area. Any ideas? Ive attatched a screen shot of the mesh in the middle of the refinement box. Thansk, Mark I also get this error in the log.topoSet: Quote:
Thanks, Mark |
||
![]() |
![]() |
![]() |
![]() |
#38 |
Senior Member
Saideep
Join Date: Apr 2015
Location: INDIA
Posts: 203
Rep Power: 12 ![]() |
Hi mark;
Well I see no attachment but, I suspect two reasons for that!! 1. Your locationInMesh value along (x y z) position. So, you generally specify the co-ordinate that you want to retain in this function. And one more thing, when you specify your locationInMesh values, you should be careful that the co-ordinate that you mention has some connection with the other region of the .stl. Else(if you have a closed portion and you mention a co-ordinate that is within the close portion of .stl) only certain portion of the point specified characteristic region is obtained. 2. If that is fine, the second reason could be that your grid refinement is not fine enough to remove or separate your hexahedral polymesh from your .stl image. Saideep |
|
![]() |
![]() |
![]() |
![]() |
#39 |
New Member
John Smith
Join Date: Jun 2015
Posts: 4
Rep Power: 11 ![]() |
Thanks again Saideep,
After changing the locationinmesh value, I am almost tertian that the problem is with the grid refinement. Which values would you recommend altering? Something like the expansionRatio or finalLaterThickness or minThickness? Thanks, Mark |
|
![]() |
![]() |
![]() |
![]() |
#40 |
Senior Member
Saideep
Join Date: Apr 2015
Location: INDIA
Posts: 203
Rep Power: 12 ![]() |
Hi there;
Good to hear that your getting close to what you need. Well what I basically do is to just change in the blockMeshDict file. Just increase the number of cells along the domain. May be double it or so.... Saideep |
|
![]() |
![]() |
![]() |
|
|
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
[snappyHexMesh] Run snappyHexMesh steps (castellated, snap, addlayer) separately | Glc | OpenFOAM Meshing & Mesh Conversion | 6 | October 8, 2021 02:50 |
[snappyHexMesh] only snap edges of specific region with snappyHexMesh | mike.franky | OpenFOAM Meshing & Mesh Conversion | 0 | July 19, 2017 06:56 |
[snappyHexMesh] snappyHexMesh. irregular cells on snap stage | Svensen | OpenFOAM Meshing & Mesh Conversion | 0 | April 3, 2015 03:12 |
[snappyHexMesh] SnappyHexMesh refine but does not snap | malaboss | OpenFOAM Meshing & Mesh Conversion | 6 | December 10, 2014 05:31 |
[snappyHexMesh] snappyHexMesh failure to snap to geometry | Yadasol | OpenFOAM Meshing & Mesh Conversion | 1 | November 17, 2014 05:00 |