CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Mesh Generation & Pre-Processing (
-   -   Orthogonal Quality Issues (

simondpowell February 19, 2013 09:46

Orthogonal Quality Issues
Hi everyone, I'm pretty new to CFD and I've waited a while to get to grips with it before posting on here but I'm having real problems with my mesh in terms of orthogonal quality so could do with some help.
I am trying to use a boundary layer using first layer distance that I calculated from a y+ value to be 0.002ft and the only way I can make orthogonal quality anywhere near a reasonable value is by making the mesh around this boundary layer really fine which leads to a massive number of elements.
Is there a better way to go about this problem? Any other information required I can give I just didn't want to overcomplicate the situation.

vangelis February 20, 2013 17:35

Hi Simon,

What meshing tool are you using?
For which solver are you creating the mesh, OpenFOAM?
What is the surface mesh element length?
Do you have sharp corners in the model?

Indeed a combination of sharp edges (>90deg) and
a high aspect ratio of prisms will lead to problems in
OpenFOAM orthogonality or skewness for example by definition,
but it would be interesting if you could give us more details.



jchawner February 24, 2013 17:21

Simon: While your question lacks important details about the type of mesh you're generating (structured or unstructured) I'll guess that the think your mesh lacks is a gradual coarsening as you move away from the wall and the tight spacing to your farfield. The manner in which you achieve that depends on your mesh type and the meshing software you're using as Vangelis stated above.

Best Regards

simondpowell February 24, 2013 17:30

Right so here is some more information about the problem I am tackling:
I am using ANSYS (FLUENT, ICEM) to investigate a validation model used to represent air flow around a generic frigate shape with geometry as follows:

When I first posted I thought that the problems with poor orthogonal quality were causing a viscosity ratio limit of 10^5, I understand for external flows this figure should be around 0.1-0.2 so this is way off. I have now reduced the minimum orthogonal quality to approx 0.22 and still get the error occurring after about 50 iterations. I appreciate this post has deviated from the initial problem and so if a moderator feels the need to move it to a different forum that would be fine.

Since I made this improvement I am now suspecting the problem lies in my boundary conditions. These are as follows:
Symmetry axis down centreline of geometry
Stationary no-slip wall for the frigate
Water surface is a zero shear stationary wall
Velocity inlet with specified velocity and gauge pressure set to atmospheric
Pressure outlet with gauge pressure set to atmospheric and backflow direction normal to boundary

I am using the k-epsilon turbulence model by means of specifying turbulence intensity and length scale as 1% and 1ft (according to other published works)

The viscosity ratio limit appears to be affecting the cells on the frigate body. I have used an inflation layer using first cell distance as 0.002ft which I calculated to satisfy my y+ value to be 30.

Hopefully I've made a rookie mistake somewhere that someone can point out easily enough. Please help me!



All times are GMT -4. The time now is 03:56.