[zuheyr]

Dear colleagues,



ICEM Hexa 2D meshing, I have a domain split into a number of blocks, utilizing construction lines which are not real physical boundary condition boundaries, just internal interblock lines, clearly.

Do I need to associate these internal curves to the internal block boundary edges?

I do need apparently to associate the construction points, which are often the end points of the construction curves to the block vertices to shape the blocks.


I am having problems when I write out the unstructured Fluent mesh as these natural boundaries are taken as walls (the default). I am overlooking something fundamental and simple. Can you please help me about this?


In an ideal world, I am hoping the smoother will smooth out these jagged line block boundaries and will end up a beautiful body-fitted like mesh... How to tell this to ICEM?


I am not using Fluent, so I need a regular quadrilateral mesh.



Thank you for reading.



Best wishes, Zuheyr

[h1773]

You need to go through some of the basic tutorials and terminologies used in the icem cfd. Tutorials are available on youtube..you can go through it. The lines between blocks are edges of the block you do not need to associte it with the geometry until you want your mesh in multiple zones in the solver.
as s example...nozzle fitted on tank.
You can create single zone mesh for tank and nozzle. Or the multizone mesh which consists of two zones one for tank and on for nozzle.
Upload images for better clarity of your doubt.

[zuheyr]

Thank you so much.



My case is a bit different, Please see the image and the link for detailed description

https://www.astro.princeton.edu/~jst...e/Implode.html


There is one fluid domain in a square of 30 cm sides in xy plane, but like a shock tube, there is a low pressure and density zone in the central area where y+x < 15cm, simply a 45 deg rotated 15 cm side square inside this 30 cm square.


Unlike shock tube, there is no diaphragm, I need to have no inner boundaries but must have 2 fluid zones like driver and driven BUT only for initialization, no physical boundary.

Of course with all symmetry boundary conditions it is better to solve to 1/4 of the geometry.

I do not manage to make a good blocking without creating internal vertices.

Thank you for reading and for help. Best wishes.