CFD Online Discussion Forums

CFD Online Discussion Forums (
-   ANSYS Meshing & Geometry (
-   -   [TGrid] PRISM layer (

sarav February 17, 2010 10:49

PRISM layer
I've a problem generating the prism-layer with Tgrid over a 3D wing: the cell-skewness on the trailing edge is very high (more than 0.99).
Who can I solve this problem? which settings do you advise me?

udayrg February 18, 2010 00:28

hi sarav

Make sure your surface mesh quality is around 0.6 to 0.65 skewness.
And then try to generate prism layers.


sarav February 18, 2010 03:55

It's impossible for me to obtain max 0.65 for surface skewness because wing trailing edge is very very peaked...
Do you think I will have convergence problems if I try to solve the case with similar volume-cells skewness?

udayrg February 18, 2010 04:39

Hi sarav

Yes certainly as your flow will be compressible flow. And flow over the wing are transonic kind of flows.


sarav February 18, 2010 05:37

I checked the surface mesh, cells on the trailing edge have skewness less than 0.6. Which could be my problem so?

udayrg February 18, 2010 06:35

Hi sarav
Check the worst elements those are in the prism by display cells quality.
Prior doing it it plot a histogram of cell skewness between 0.9 to 1 with 10 equal parts.

And then i would like to know how are you managing the trailing edge prism nodes.

In other words the the prisms traveling from pr. side and suction side meet at the trailing edge how are you modeling the prisms there.


sarav February 18, 2010 07:21

Hi uday.
Well I've done the plots.
I find in the fluid zone only one cell over 0.99. Instead in prism-cells I find zero cell over 0.97. I set the two plots as attachments.

I'm managing the prism over the trailing edge as the rest of the wing...that is to say, for example in my last try, with 10 layers with linear growth of 0.1 of slope and first height of 0.5.

Can I menage the prism in a particular way only on the edge?



udayrg February 18, 2010 07:36

Hi sarav send me your mesh file

My mail id is

I will look into it and suggest you and to some extend try to finish the meshing may be upto prisms.


sarav February 18, 2010 08:06

1 Attachment(s)
Hy uday,
sorry but I can't send you the geometry.
I set as attachment a view of the problematic cells, hoping that you could understand the problem from this view.

udayrg February 18, 2010 08:28

Okay sarav

1)Refine the surface mesh at the trailing edge region before prism generation
2)I would suggest to put a very small surface at the trailing edge instead of edge. i.e. a very negligible width plane surface and mesh it minimum one element. And then you wont be having any problems generating prism layers on it.

I am afraid that the solution (1) alone. It not difficult however you will have to try number of times with grid refinement at the trailing edge. and You can play with shrinkage option for prism layers.


sarav February 18, 2010 09:28

Ok I try to refine the surface mesh, and eventually I will modify the geometry in future...
For shrinkage options do you mean "keep first layer offset", gap factor and so on?
thanks again, you are very kind!

wadebogs March 3, 2010 17:24

Try "minimum height" growth method instead of "uniform"

Use Mesh -> Tools -> Auto node move after creating the prisms to locally smooth bad cells.

dvb018 September 9, 2010 10:57

Even i am facing same problem while generating prisms & my surface skewness is 0.64 (max). What els can be done to reduce the skewness?? I am using the last ratio method in growth options.

All times are GMT -4. The time now is 14:48.