[happy]

Hi ghorrocks and every one

This tutorial 12: Room with Air Conditioning in this link below
http://ifile.it/3vcyb1a give us how we create geometry of room....
my question is why it is ignored thichness of that room? for my case when I considered the thickness of walls and floor, I have problem when I'm going to imprint faces and select 2D region.
My case is explained in this tread
http://www.cfd-online.com/Forums/cfx...-geometry.html
My problem related to imprint faces is that Imprinted faces are only in one side...?
By the way I work with ANSYS 12.0 and try to draw my geometry by CFX-workbench....another question, How I can make sure that my geometry correct? by correct meshing or what.......

Please, help me and tell me if I took the right way to draw my geometry?

Regards

[happy]

pls,,,help

[Darren Leong]

Hi Happy,
Your problem definition's going all over the place XD! Work with one post for starters.
e.g. [CFX] Tornado Chamber modelling. Help pls

For simplication, you only need to model the wall surfaces as i assuming you're investigating on only the airflow and not energy loss through wall. I'm not sure of the approach you've taken but try slicing with zero thickness instead of imprint.

Ps. You should be having some in-house support for this kind of project. Is this for academic or industry?

-dazza

[happy]

Actually, It is academic...Could U, PLZ, give me more explaination..I need to consider the thickness for my chamber....
I wait

[Darren Leong]

I'll need more information. Is it possible to post ur project plan or ur methodology at least.

Few questions:
1. Are you carrying out the experimental work too?
2. Clarify how "thickness" affects your setup rather than us assuming/giving a solution for you.

Re: Model only the wall surface (means there's no mesh inside the wall domain itself). Put an inflation layer around the wall boundaries in the air domain. If you can follow this then don't worry bout the questions above.

Let me know where you're at (e.g. geometry, meshing, pre-solver, post-solver or reiterating the process) Will provide more guidance if you do need it =)

[happy]

Hi again

-believe me or not I have not plane, what I think now how I draw my geometry and at the same time I need to consider the thickness.
-Of course, I will do practical work but is not at thie stage
-thickness will affect espicially at floor( foundation) where there are heat source is with contact to floor.it means that some of heat will go through foundation( loss heat by coduction).
-of course not, I'm going to consider boundary conditions and I have fluid domain & solis domain......it is not only wall surface's model.

I wait your reply....

[Darren Leong]

Assuming you have 2cm thick walls and floor, model them first. Create your air domain of the same size and then use the subtract boolean. This will leave you with just the air domain with the wall geometry as the boundaries (this should eliminate the need for you to imprint faces ^_^) . Will leave the meshing to you (post ur mesh model and we'll work from there how to refine it).

In ur cfx-pre:
1. Specify the wall boundaries as smooth wall (if you do have the surface roughness from an exprimental setup, u can specify that instead)

ps. (couldn't wait XD) ur wall thickness should only affect ur flow regime at the inlet if you're not considerig heat loss

[Darren Leong]

Posted above the same time as ur reply.

---
Just to redefine the above for ur situation, in the subtract boolean tick preserve bodies. That'll leave you with your air domain, and solid domains for ur walls and floor =) Sketch a box line for ur heat source and extend the profile on the floor using slice with zero wall thickness. Name the selections and you should be able to start meshing from there.

You might want to look into ICEM for your application but CFX will work fine in expense of computational time and resources.

Ps.
1. You can post ur geometry model if you need feedback. There shouldn't be any geometry error since it's all flat surfaces and you're working with ansys DM. You can check to be sure in DM by using "fault detection" under Tools>Analysis, and use repair features if needed under Tools>Repair.
2. There's a trick to consider too. U'll have to combine the domains as one part (e.g. as sub-domains). This allows CFX to mesh the interfaces identical on both sides, allowing you to define the interface as 1:1 later in CFX-pre.

[happy]

Hi again...
Thanks a lot for ur fantastic advices

• where I can find (subtract boolean). does it mean thin/surface button? If so, I used but with different way.
• What U mean with ( XD ).
• I have tried to make my geometry more than 4 times and I have some errors after meshing...
• How I can set domains before I make my meshing as U said.What I know that set the domains is after create the mesh( at CFX-pre stage).
• this is my geometry from many sides, but this is without your recommindation
• finally...could, U, pleasssssse help me more to create my geometry

Best regards

[Darren Leong]

Sry, Happy. You'll have to forgive me for finding this slightly amusing. The "XD" is an emoticon - eye squint + big smile, similiar to a facepalm but more laughter in t.

Back to work:
1.The boolean function's under the "create" tab in DM. Search the help menu in ansys on how to use it.
2. -
3. Post a pic of ur geometry as it is. Will look at it and get back to you later 2night or 2moro.
4. Work on get the domains first and we'll go from there. You still be able to mesh without having them in one part.
5. where XD?
6. I could but i have to highlight now that I can only advice on how to get the setup right and not on how to use Ansys. Took me a month to get comfortable with the package and it's worth going through it urself. You'll need to understand what ansys does by going thru the help doc first when u bump into a problem. Will still guide you where i can tho I don't have details of your project.

CFD solutions are highly dependent on the user's input. What you get from it and is what you put in it. Remember it's only a tool, u're the engineer/hyrodynamicist....or will be =)

Best wishes,
Dazza

Ps. What level of research (ug, ms or phd) are you doing? If u have a supervisor, best is to start poking for directions and training if needed. You should have a project plan to work with at start if it's ms n above; it'll cover you objectives, methodology and lit review. Don't leave this part too late or u'll be wasting time doing unguided work.

[Darren Leong]

Geometry looks good =) The only complicated part will be meshing the heat source cylinder later.

1. For the rectangles (inlet), you'll need to remove so it'll be a hole. That'll leave you with four holes on side as the inlet.
2. Create a box with the exact outer dimension (its position superimposed exactly on ur present model)
3. Subtract using boolean as mentioned earlier (if done right, u'll have 2 domains only). Post pic with ur tree on the left; printscrn.
4. Will advice further once 3 is done. You can choose to partition the domains later etc. In the meantime, read up on CFD. A good book to start is Computational Fluid Dynamics: A Practical Approach by Jiyuan Tu, Guan Heng Yeoh Chaoqun Liu 2008. A better start is go through CFD-wiki =)

[happy]

Helloagain.......

Thanks a lot for ur helping my dear...
U now can see ur email to revise my geometry...
U know second time I create my geometry I cut the rectangler faces but when I want to set BCs I cannot select them by mouse in CFX-mesh kind...

[Darren Leong]

Hi Happy,

Don't progress to CFX-mesh yet til you have your domains setup. Will wait for your print screen of the tree and domains =)

Another thing, is it ok if u continue to post ur pics and send only the actual work files thru pm as you have done if there isn't an intellectual property agreement on ur project. It would be good as reference for other ppl in future ....n I'm pretty bad at checking my pm XD

[happy]

Hi again....

I will apply ur advices but U have to check ur email
Thanks so much for your source..it is amazing
Regards.