Advice needed for meshing multiple channels

[ukok]

Hello all,

I am meshing multiple channels with a box in front, flow from left to right. I'd like to know if there is better way to mesh this fluid domain. Ideally, I'd like to two blocks for the box and channels. but have no idea how to do it, and join it in code-saturne.

Any advice will be appreciated.

Thanks

Sam

[Brian Angel]

Hello Sam,

Are the channels regular in their shape, i.e. circular or rectangular? Are you looking to use open source or commercial meshing codes? What cell type(s) are you looking to use: tet, hex, poly, a combination? Do you want a conformal or non-conformal mesh?

Regards,

Brian Angel.

[ukok]

Hello Sam,

Are the channels regular in their shape, i.e. circular or rectangular? Are you looking to use open source or commercial meshing codes? What cell type(s) are you looking to use: tet, hex, poly, a combination? Do you want a conformal or non-conformal mesh?

Regards,

Brian Angel.

Dear Brian,

Thanks for your reply.

All the channel are regular in the 2D shapes, less 1mm in thickness and 500mm long. I am open to all available options (open or commercial) since I am in the MU. Regarding cell type, I prefer the hex since I need the boundary layer information in the thin channels. My current mesh 0.25M notes is working but slow.

I am new to the conformal mesh. but if it could help, I would like to try it.

Any suggestions are welcome!

Many thanks

Sam

[Design_BIB]

Hello Sam,

In my opinion, There are 2 simple methods in this manner. Please following below ...

1) Divide each "Box and Channel Layer" by a "Red Line".




2) Divide the Box and Channel by a "Red Line". It will separate them into 2 parts.





And you can couple it with "Max Size" in 1-D Algorithm of Wire Discretization. If the max size of both parts are matching, the mesh will be fine. I cannot tell you a suitable max size for this geometry. I know you can find it by yourself easier.


Hope this helpful


Best Regards,
Chayawat

[Brian Angel]

Hello Sam,

Following on from the reply from Chayawat and given the relative simplicity of the geometry and the length and height of the 2D channels, it seems to me that the meshing process that you are currently using is adapted to the geometry.

I would ensure that the mesh is conformal at the interface between the main block on the left and the channels on the right. Also, is your flow laminar or turbulent? If laminar, you will need to mesh the near wall regions very finely in order to resolve the boundary layer. If the latter, please pay attention to the y+ requirements of the turbulence model that you are using.

Regards,

Brian Angel.

[ukok]

Hello Chayawat and Brian,

Thanks for your replies.

That is exactly what I did for my mesh, and instead of using "max size", I defined a density distribution to make more grid close to the wall of thin channels.

Does anyone know the different between constant time step and variable time step option? I tried a few test with small number of iterations. they are different.

In the variable time step, which are the main control parameter? time step or max CFL number?



Thanks
Sam

[Design_BIB]

Hello Sam,

Sorry for cant help you too much.
But I want to know something about Time-step as same as your last question.
Please let me following this topic with you too.


Best Regards,
Chayawat