[abhi084]

Picture1.png

Picture2.jpg

hi,
i'm having problem generating a mesh in the drawing attached above.
Although i am able to generate a structured mesh here but wen i use inflation then my grid is not remaining structured (not fully hexahedrals). I am getting tets and hex's. How to solve this problem. And also can we use inflation in 2d. Please advise on how to go about meshing this domain.

Thanks in advance

[mvoss]

Create the mesh by applying an extruded 2D-Mesh (for CFX) for the body, select the appropriate faces for target and source manually. Then pic the edges at the source and apply the sizing-functions by hand for every ensemble of edges.

[abhi084]

thanks for that. infact thats what i was doing by giving individual sizing to each edge, but the problem arises when i try to put inflation in the body.
My main problem occurs as i mentioned above is due to inflation. how to make a BL resolution by keeping the full domain with structured hex's
Plz help in that regard.

[mvoss]

You need to create the "inflation" from the line-segments.

[abhi084]

okay so u mean just divide it into sub domains and then use edge sizing and the biasing?

[huey1080]

When you are dealing with steps in ICEM and boundary layer, the best way is to go with 45 degrees split. That way you have a perfect connection between the faces of your boundary layer.
Look through the snapshots : using split-block, delete-block, merge-vertices

[amin.z]

Originally Posted by abhi084

thanks for that. infact thats what i was doing by giving individual sizing to each edge, but the problem arises when i try to put inflation in the body.
My main problem occurs as i mentioned above is due to inflation. how to make a BL resolution by keeping the full domain with structured hex's
Plz help in that regard.

Based on my experience, unfortunately, in ANSYS meshing we aren't able to create inflation for mapped mesh!

[abhi084]

yes i tried the same earlier but we cannot do mapped meshing with inflation. It gives it as an invalid method. What i think is the problem is that inflation is used basically for generation of prism layers below the tets and hence can be basically used only with unstructured meshes.
based on above replies i guess the only way left is to either mesh in ICEM as suggested or the best feasible method which i can see is by using bias in edge sizing, but how to control then first node distance.....any help wud be appreciated.....i saw some other threads related to it, but what i came across is that there is no direct way for giving growth ratio,no. of layers etc. in bias so what i think is i will have to divide the BL in a separate block and then mesh seperately......Is that the correct way to do?

[mvoss]

You could add extra "slices" above the walls and generate defined ratios from this "extra-geometry"/ new edges.

[abhi084]

And yes what about making the mesh in ICEM. I did try some trials of other geometries some time back. There we have bunching laws so wouldn't it better for BL meshing as we would have greater control there. Also we can make structured hexa's there too ryt....so finally which would be better (have yet to try in ICEM) ansys meshing or ICEM.

[mvoss]

No doubt - ICEM.