[liadpaskin]

Dear all,

I am performing DES computations with focus on tip vortex flows. Being a hybrid RANS/LES approach it is a key point to have a fine control over the mesh sizes and behaviors so that after some research into ICEM cfd capabilities i've chosen a multi-zone approach for the meshing.

The model's geometry is a 3d propeller blade on a rectangular channel. Previous computations gave me the vortex trajectory, from which i've drawn two refinement cylinders that i input together with the geometry. A refinement box is also included.

geometry01.jpg

geometry03.jpg

The plan is to build hexa meshes on the cilinders and outside of the refinement box (farfield). Between the outer cilynder and the box tetra would take place.

-First of all, this is the first time: i input geometry for mesh control; and work with blocking. Does this plan sounds good?

If so... i follow to the specific issue:

2d blocking as either swept or free, works fine, but no matter what the operation 2d to 3d excludes the regions i am more interested in, i.e., the cylinders.

2dblock.jpg

mesh04.jpg

mesh05.jpg

Most of the time it returns one block with the whole domain. Excluding outer surfaces, or introducing cut planes brings the block closer to the region of interest most of the time ignoring the cylinders but capturing the blade. I was hoping that limiting the initial input i could get something i could manufacture into the final blocking.

The best i could do until now is to block only a half (y>0) of the refinement box, capturing the outer cylinder and the body. I hoped i could build the blocking from this one but: it seems quite laborious to add the inner cylinder; my problem has no plane of symmetry; and mirror this block doesn't work.

- How should my procedure differ from the ones exposed at https://www.ansys-blog.com/ansys-icem-cfd-multizone/ or https://www.youtube.com/watch?v=HmEzZ1zZnsU ? Any other reference in mind?

Other general questions regarding the procedure:

- One fluid one body? Should i have multiple material points for one body than? Do i need an ORFN body?

- Will i be able in the end, to exclude the interfaces as boundary conditions for export?

I am having so much trouble with the blocking that it feels i'm doing something really wrong. So i appreciate any kind of help. Thanks in advance.

Best regards,
Liad.