generating OpenFOAM meshes with pointwise

robo · April 29, 2014, 13:39

Fairly new to openfoam. I've been working on a ship hydrodynamics problem and having a lot of issues. I have setup, run and validated the results from a simple model for steady free surface flow using LTSInterFoam on the surface piercing NACA 0024 experiment. Everything went fairly smoothly. However when I try to run using my actual ship geometry, everything goes to hell.

I can run both my validation case as well as the supplied DTCHull tutorial problem using the solver settings that I validated, so I think that the issue must be my mesh. When I run checkMesh I get a lot of skew cell's and non-orthogonal failures. I generate the mesh in pointwise, however when I run the normal check's I would use in pointwise everything comes up fine. If I run the mesh in other CFD software everything comes up fine. I can only think that there must be some export issue between pointwise and openFoam, or that openFoam is MUCH more sensitive to meshing issues then other solvers.

I would rather not have to use snappy, as I find it a huge pain in the a** to get boundary layer resolution in snappy, and my my company already pays for pointwise which I can use fairly well. Does anyone have any recommendations for using meshes generated in pointwise in openfoam solvers?

robo · April 29, 2014, 13:47

I should add that I also get a ton of failures from checkMesh due to aspectRatio, however I am really unclear on how I could NOT get those failure on a boundary layer mesh... It seems completely unreasonable to require that the surface be resolved down to a similiar resolution as the wall normal direction. My ship is ~120 m long, running @ 5 knots, giving Re ~ 530e8, @ Y+ = 100 that would be somewhere around 1.0e-3. I am trying to resolve the surface @ between 0.5 and 0.1 m, giving an aspect ratio between 100 and 500.

April 29, 2014, 13:39	generating OpenFOAM meshes with pointwise	#1
robo Member robo Join Date: May 2013 Posts: 47 Rep Power: 12	Fairly new to openfoam. I've been working on a ship hydrodynamics problem and having a lot of issues. I have setup, run and validated the results from a simple model for steady free surface flow using LTSInterFoam on the surface piercing NACA 0024 experiment. Everything went fairly smoothly. However when I try to run using my actual ship geometry, everything goes to hell. I can run both my validation case as well as the supplied DTCHull tutorial problem using the solver settings that I validated, so I think that the issue must be my mesh. When I run checkMesh I get a lot of skew cell's and non-orthogonal failures. I generate the mesh in pointwise, however when I run the normal check's I would use in pointwise everything comes up fine. If I run the mesh in other CFD software everything comes up fine. I can only think that there must be some export issue between pointwise and openFoam, or that openFoam is MUCH more sensitive to meshing issues then other solvers. I would rather not have to use snappy, as I find it a huge pain in the a** to get boundary layer resolution in snappy, and my my company already pays for pointwise which I can use fairly well. Does anyone have any recommendations for using meshes generated in pointwise in openfoam solvers?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Tool to convert OpenFoam mesh to SU2 mesh (2D and 3D)	Combas	SU2	44	March 12, 2024 00:35
[Other] Pointwise -> OpenFOAM, Axisymmetric	Nucleophobe	OpenFOAM Meshing & Mesh Conversion	2	May 22, 2015 05:19
[Other] From Pointwise to OpenFoam 2.1 cyclicAMI...	lordvon	OpenFOAM Meshing & Mesh Conversion	9	February 9, 2015 15:08
OpenFOAM Foundation Releases OpenFOAM v2.3.0	opencfd	OpenFOAM Announcements from OpenFOAM Foundation	3	December 23, 2014 03:43
The OpenFOAM extensions project	mbeaudoin	OpenFOAM	16	October 9, 2007 09:33

April 29, 2014, 13:47		#2
robo Member robo Join Date: May 2013 Posts: 47 Rep Power: 12	I should add that I also get a ton of failures from checkMesh due to aspectRatio, however I am really unclear on how I could NOT get those failure on a boundary layer mesh... It seems completely unreasonable to require that the surface be resolved down to a similiar resolution as the wall normal direction. My ship is ~120 m long, running @ 5 knots, giving Re ~ 530e8, @ Y+ = 100 that would be somewhere around 1.0e-3. I am trying to resolve the surface @ between 0.5 and 0.1 m, giving an aspect ratio between 100 and 500.