Learning Forum

[cheolmae]

The models I work with have a volume ranging from as little as 10 to as many as 15,000. In the case of Named Selection, up to 100 or more have been specified. And since the model is so big, I split it into 3 or 4 pieces and repeat it 3 or 4 times.

Workbench Meshing always returns to 1core. It's too slow.
Also, Cutcell can no longer be used.
Meshing and Named Selection in Fluent are a pain in itself for me.

Please combine Fluent Meshing's Multicore Meshing with Workbench Meshing.
Alternatively, please export a 3D modeling file with only Named Selection and Local Sizing set without creating a mesh in Workbench Meshing, so that meshing can be done in Fluent Meshing without any special settings (actually, this is my core request).

Please don't tell me to do Named Selection in Space Claim.
The Named Selection process in Space Claim is also painful.
Also, Space Claim cannot properly share topology the 3D model I created.

Also, don't ask Design Modeler to do Named Selection.
The work speed is horribly slower than in Workbench Meshing, and the GUI becomes horribly slow after Share Topology, making it virtually impossible to work with.
Also, Named Selection made in Design Modeler cannot be read properly in Fluent Meshing and becomes fragmented, resulting in a very large number. Moreover, the response speed is so slow that it is virtually impossible to work with.

Lastly, I would like to ask you again.
Please export a 3D modeling file with Named Selection and Local Sizing setting information without creating a mesh in Workbench Meshing, so that meshing can be done in Fluent Meshing without any special settings.

p.s. Please know that many people are still using Design Modeler.

[Rob]

We're aware many still use DM - I use a mix of Discovery and DM depending on the need. 

Ansys Meshing was designed for Mechanical, and is part of Mechanical in their interface. When it replaced GAMBIT many of the issues you raise were noted. Most have been resolved, but this is why TGrid was further improved and later renamed Fluent Meshing. 

Ansys Meshing typically creates a volume mesh that's then projected back to the surfaces, so there isn't a "surface mesh" in the classic sense. Fluent Meshing in parallel isn't as efficient as the solver for parallel meshing, I rarely run more than 2 cores on Fluent Meshing regardless of what I'm doing. I run 96+ cores on the cluster as needed. 

On the public forum I can only point you at the Learning content for the various workflows. If you have commercial software please talk to your local team: there are a lot of tricks to assist in the geometry and meshing stages and some may be very specific to your tasks. 

[cheolmae]

I am a commercial user, and my company pays maintenance fees every year. So, of course, I received technical consultation. However, as a result of the consultation, the person in charge responded that this is difficult for them as well.

Fluent meshing is good even if the parallel CPU is inefficient.

Workbench Meshing must be run on a single core.
I have a 128 Core Workstation and a separate 24 Core Workstation for work. The CPU can not work before mesh is finished.
There are sufficient CPU resources available for Mesh.

And there is a more important point. That's because Ansys has made almost no updates for CFD in Workbench meshing, and there are no plans to do so.

So, we want to use Fluent Mehsing.

Lastly, I would like to ask you again.
Please export a 3D modeling file with Named Selection and Local Sizing setting information without creating a mesh in Workbench Meshing, so that meshing can be done in Fluent Meshing without any special settings.

[Rob]

I'll put an enhancement request in. 

[cheolmae]

Thank you very much!

There seems to be an error in Google Translation, so I'd like to explain once again.

Please can make the mesh with the file what have a 3D modeling data with Named Selection and Local Sizing setting information worked in Workbench meshing using Fluent meshing.

 

 

[Rob]

You're welcome. 

That's roughly what I put into the system. As an aside, if you contact your support team they can also do this and provide a tracking ID alongside tagging your userID against the request. I'm not permitted to do the latter on here as I can't confirm your user credentials as the Forum is "external" and "public" from a Legal definition.