mesh quality statistics

Altair Forum User

I have got a question, concerning the mesh quality. Can you please tell me, what these values, from the MeshSim.txt file are standing for?

ydigit

What are the mesh quality statistics at the end of the Meshing output file?

SKEWNESS: 
The equilateral or equivolume skewness is non-dimensional ranging from 0 (perfect tetrahedron) to 1 (degenerate). 
S = ( V_opt - V ) / V_opt 
V_opt = 8/27 * sqrt(3) * R^3 = Perfect tetrahedron volume 
V = tetrahedron volume 
R = circumsphere radius

ASPECT RATIO: 
Aspect ratio is non-dimensional ranging from 1 (perfect tetrahedron) to infinity (degenerate). 
AR = Longest Edge / Shortest Altitude

DIHEDRAL ANGLE: 
Cosine of the dihedral angle of the tetrahedron. Smallest and largest are given.

rByR: 
The unitized ratio of the insphere radius to the circumsphere radius of the tetrahedron. You see '-1.0' as the minimum if there are non-tetrahedral elements.

Altair Forum User

The question is how to reduce the max. volume skewness?

I would be glad for it.

Thanks

ydigit

The typical reason for high value of volume skewness is boundary layer elements. In order to reduce it, you would have to mesh the surface finer (will increase the total element count) or use coarser BL elements (has impact on accuracy!).

 

Altair Forum User

Thanks for your suggestion, ydigit! 

Altair Forum User

Hi again,

 

i am still thinking about the reduction of max volume skewness. To get a good y+ value, you have to create a first layer thickness at about 10e-4, 10e-5 m. Or finer or coarser depends on your case. So, in this case i could probably have really fine mesh on all surfaces (advice from ydigit) that is not always possible, if you have big models with details. For Hypermesh, there is a option for the target volume skewness. At acusolve, it was for me also possible to reduce the volume skewness with option volume improver shape metric. But without boundary layers! 

What is the way to get a good mesh with a volume skewness 0.95?

Thanks for help

Best regards   

acupro

AcuSolve is much more tolerant of element skewness than most other solvers, so in most cases there is no need to focus so much on individual element quality, even with tetrahedral boundary layer elements.  So 'good mesh' is related more to the locations of nodes making sure the gradients are captured, rather than how those nodes are used to define elements.