Boundary layer grid for Wigley hull by snappyHexMesh
Dear all,
Hello, I'm working on grid generation around Wigley hull using snappyHexMesh. The code version and working environment is OpenFOAM 2.0.1 under Ubuntu 10.04LTS (64bit). The case I'm trying to modify is: /tutorials/multiphase/LTSInterFoam/wigleyHull For the default run (i.e. don't modify anything written in "snappyHexMeshDict" in the tutorial directory), I have no problem and can view mesh by paraFoam. When I tried to insert boundary layer mesh by using "addLayers" as well as modifying "addLayersControl" parameters , mesh has been created in 4 steps, i.e. \0: Mesh by blockMeshDict \1: Before snap \2: After snap \3: After inserting boundary layer mesh In log.snappyHexMesh, I don't see any error message and checkMesh tool doesn't show error message neither. Yet the mesh in \2 and \3 looks identical, in other words, no boundary layer mesh is inserted. Has anyone encountered such problem? Any comments to figure out this problem will be greatly appreciated. I'm also ready to upload my snappyHexMeshDict as well as relevant files to somewhere if necessary. Thanks in advance for your help. Best wishes, Novak |
Hello Novak,
it would be nice to see some pictures or even your SHM dict. CU |
Hi CU,
Thanks for your offer. In consequence, the problem was resolved. I appled wrong patch name under "nSurfaceLayers". Once the right patch name was applied (e.g. in the Wigley hull example in OF2.0.1, it is "hull_wall" and NOT "hull") I managed to insert boudary layer mesh. Best wishes, Novak |
Hi Novak,
Did you manage to obtain a nice BL mesh all round the hull (especially at body ends)? If yes, I'd be curious to see your sHM.dict file :) Regards, Ben |
2 Attachment(s)
Hello Ben,
Thanks for your comment, actually that's currently my next issue to solve. Attached figure is the example of the boundary layer (at midship yz-plane) I created using the SHMDict as per also attached. As seen in the figure, the mesh at z-top boundary look collapse even though checkMesh command doesn't give any error message. This is the 1st issue to be resolved, and I'm working on the geometry for double-model flow. In such case, I don't have to put refinement boxes on free-surface. Second issue is, when I try to create boundary layer mesh with minimum spacing equal to 2e-6 (i.e. for mesh without wall function at Rn=1e6 simulation), SHM does not even insert any boundary layer mesh. The density of background mesh generated by blockMesh seems not a problem. Anyways, I'll keep working on this problem. Sorry in advance that my progress may be quite slow, but I'll keep checking thread. Thanks. Best, Novak |
Interrupted layers
3 Attachment(s)
Hey Novak,
you might have discovered the posts in this forum stating that the "collapsing" of layers at the boundary stops if you set the boundary type to empty. This worked for me. Unfortunately my layers look fine in the midship area but unaccaptable everywhere else. Do you have a fix for this problem? Or does anybody else? Regards, Friederike |
Hi all,
I found the following things: -Increase the number between cell refinement layers (control to be found in sHM dict). This increases the general quality of the mesh -Run sHM for each option seperately -Setting a patch to empty doesn't always work; especially NOT when the object (hull) is not perpendicular to the patch which is set to empty -I think that it's also good to consider to run the layers within sHM layer by layer. Still it doesn't give you the most beautiful mesh but I guess it's the best you can get. Good luck! Ralph |
Possible clue to BL problems
3 Attachment(s)
Hi,
I've been working on the Wigley hull tutorial and have been frustrated by the BL grid. I have first run the tutorial as it came and the grid did not look good at all. I tried several changes hoping to improve things, following advises found here on the forums, and things have improved, until I hit a wall. In the part of the wigley hull with the vertical sides I can get a beautiful BL grid, and so if I look at a midship section, but at other waterline or sections things don't always look good. I think I tracked down the cause, but don't have a solution. In my case the BL grid is not grown where the surface grid on the hull has "weird" looking shaped faces that are due to the snapping. These not-four-sided faces are the results of the snapping when it switches from one row in the castellated mesh to the next. It doesn't always happen, as sometimes the snapped face is still topologically a rectangle, but in a lot of cases that's not the case. And those location seem to correspond to the locations where the BL doesn't grow properly. (see the attached pictures). I thought that maybe it could be because I was using "relativeSizes true" that it was not growing the BL properly, but when I switched to "relativeSizes false" and changed the "finalLayerThickness" from 0.6 to 0.004, things actually got worse. Does anybody have a suggestion on how to solve this? snappyHexMeshDict: Code:
/*--------------------------------*- C++ -*----------------------------------*\ Code:
/*--------------------------------*- C++ -*----------------------------------*\ |
Update
3 Attachment(s)
I wanted to give an update on the addLayer grid issues I have been having.
I had been working in OpenFOAM 2.1.0, so after reading the 2.2.0 improved snappyHexMesh and in particular the addition of layers, I upgraded. And things did work a little better, but still this operation seems far from reliable. I should say I have not had the time to redo all the tests I had done under 2.1.0, but it seems that the underlying problem is still the same. The addition of layers fails where the surface grid is at a jump in row or column in the castellated mesh. It doesn't happen every time, but when it does it does happen there. I'm attaching a picture (BLatWL_detail2.png) that shows the improvement w.r.t. the post above. I used the slipFeatureAngle, and set it first to 30 then to 1, with no appreciable difference. The midship section where all the surface elements are very close to rectangles ends up having a beautiful BL grid, the bow section does not. If anybody has any suggestions on how to get a BL grid that looks like the midship one everywhere on the hull, it will be greatly appreciated. Thanks, Claudio |
All times are GMT -4. The time now is 09:17. |